• No results found

DETERMING THE MOST COST EFFICIENT DISTRIBUTION STRUCTURE : A CASE STUDY INVESTIGATING IF A CHANGE OF INBOUND DELIVERY POINT WOULD REDUCE THE TOTAL COST

N/A
N/A
Protected

Academic year: 2021

Share "DETERMING THE MOST COST EFFICIENT DISTRIBUTION STRUCTURE : A CASE STUDY INVESTIGATING IF A CHANGE OF INBOUND DELIVERY POINT WOULD REDUCE THE TOTAL COST"

Copied!
105
0
0

Loading.... (view fulltext now)

Full text

(1)

Spring 2018 | Isrn-Number: LIU-IEI-TEK-A–18/03107–SE

DETERMING THE MOST COST EFFICIENT

DISTRIBUTION STRUCTURE

- A CASE STUDY INVESTIGATING IF A CHANGE OF INBOUND DELIVERY POINT WOULD REDUCE THE TOTAL COST

Authors:

Pettersson Ida Sanfridsson Jacob Examiner:

Huge Brodin Maria

Supervisor at Linköping University: Haag Linnea

Supervisors at Toyota Material Handling: Eriksson Bergman Ulla

(2)

Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare – under 25 år från publiceringsdatum under förutsättning att inga extraordinära omständigheter uppstår. Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning. Överföring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten, säkerheten och tillgängligheten finns lösningar av teknisk och administ-rativ art. Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfattning som god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan form eller i sådant sammanhang som är kränkande för upphovsmannens litterära eller konstnärliga anseende eller egenart. För ytterligare information om Linköping University Electronic Press se förlagets hemsida http://www.ep.liu.se/.

Copyright

The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional circumstan-ces. The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/hers own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assu-re authenticity, security and accessibility. According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement. For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page: http://www.ep.liu.se/.

c

(3)

The thesis has been interesting, fun and developing. It has included many different challeng-es and it has given us an insight on how a distribution network actually works. It has also given us valuable work experience from a multinational company. We thereby want to thank TMHEL for the opportunity to write our thesis at the company.

We would like the give our sincere appreciation to our supervisors Mikael Öberg and Ulla Eriksson Bergman at TMHEL for help with data collection, support and great discussions along the way. We would also like to thank all other people at TMHEL for pleasant breakfast discussions together with demonstrated commitment for the importance of the task.

We would also like to give our our sincere appreciation to our supervisor at the university Linnea Haag together with our opponents Johan Hammers and Johan Sarwe, who have con-tributed with great feedback and ideas to make the thesis even better.

2018, Mjölby, Sweden

(4)

Toyota Material Handling Europe, TMHE, is one of the worlds leading forklift manufactu-rers, who produces both warehouse and counterbalanced forklifts at three production sites across Europe. At each of the three production sites, there are three central warehouses, CWs, which stores spare parts to the forklifts. These CWs are managed by Toyota Material Handling Europe Logistics, TMHEL. TMHEL are responsible for all CWs and regional wa-rehouses, RWs of spare parts, so that spare parts is available to service technicians in Europe within 24 hours.

The current distribution structure of the spare parts is that the items from the suppliers are first transported to one of the three CWs before they are distributed further to the RWs. The identified problem with this distribution structure is the few CWs, which sometimes leads to longer transportation routes. For example, a RW in Germany can not order items directly from a supplier in Germany. It means that the goods first needs to be transported to a CW, before it can be transported to the RW in Germany. TMHEL thereby wants to investigate if the total cost could be reduced, if the supplier could deliver directly to either a CW or a RW. This means that the possible inbound delivery points from the supplier increases from three warehouses to six. The purpose of the thesis was thereby formulated as:

”The purpose of the thesis is to investigate if a change in inbound delivery point from three suppliers of spare parts can reduce the total cost.”

The current total cost and the total cost for the five alternative inbound delivery points were thereby calculated and compared. The total cost were calculated based on a general total cost model by Oskarsson et al. (2013), which includes transportation cost, inventory carrying costs, warehousing costs, administration costs and other costs. After the total costs were calculated, a sensitivity analysis was made regarding changes in the input data, to see how changes would affect the total cost and which warehouse that would be the most cost efficient inbound delivery point.

The results in this thesis showed that all three investigated suppliers should keep the existing inbound delivery point in Mjölby, since it gives the lowest total cost. The results also showed that the transportation cost was the largest cost parameter, which also affected the results the most. The current inbound delivery point gave the lowest total cost mainly due to the low transportation cost when transporting together with the production site, and the large demand at the current inbound delivery point.

However, the sensitivity analysis showed that the most cost efficient inbound delivery point would change, for all investigated suppliers, if the input parameters changed. For one supplier, only small changes in the input data was required until another warehouse was more cost efficient to use as the inbound delivery point. For the two other investigated suppliers, large changes in the input data were required until another inbound delivery point was more cost efficient.

(5)

Table of Content

1 Introduction 1 1.1 Background . . . 1 1.2 Purpose . . . 2 1.3 Company Directives . . . 2 2 Present Situation 3 2.1 Toyota Material Handling Europe Logistics . . . 3

2.2 Present Distribution Structure . . . 5

2.3 Suppliers . . . 6

3 Frame of References 10 3.1 Spare Parts . . . 11

3.2 Modelling the Supply Chain . . . 13

3.3 Total Cost Analysis . . . 15

3.4 Localisation Methods . . . 19

3.5 Sensitivity Analysis . . . 22

4 Problem Definition 23 4.1 The Studied System . . . 23

4.2 Purpose Breakdown . . . 27

4.3 Summary of the Questions . . . 31

5 Method 32 5.1 Phase 1 - Understanding the Problem . . . 34

5.2 Phase 2 - Planning . . . 35

5.3 Phase 3 - Execution & Analysis . . . 37

5.4 Method Criticism . . . 43

6 Supplier A 47 6.1 Main Question A - Total Cost . . . 47

6.2 Main Question B - Sensitivity Analysis . . . 53

7 Supplier B 59 7.1 Main Question A - Total Cost . . . 59

7.2 Main Question B - Sensitivity Analysis . . . 65

8 Supplier C 69 8.1 Main Question A - Total Cost . . . 69

8.2 Main Question B - Sensitivity Analysis . . . 75

(6)

10 Discussion 82 10.1 Limitations . . . 82 10.2 Future work . . . 83 10.3 Recommendations . . . 85 A Appendix - Time Plan

B Appendix - Literature Search C Appendix - Total Cost

(7)

List of Figures

1 THME Organisation . . . 3

2 Overview over the system in Europe. . . 4

3 TMHEL current structure . . . 5

4 TMHEL example on new distribution structure . . . 6

5 Map of the transportation routes from supplier A . . . 7

6 Map of the transportation routes from supplier B . . . 8

7 Map of the transportation routes from supplier C . . . 9

8 Literature structure . . . 10

9 Product lifecycle and its phases . . . 11

10 Relationship between total cost and customer service . . . 14

11 How costs vary depending on the number of facilities in the supply chain . . 15

12 Ordering-point system with varying demand . . . 18

13 Density of demand . . . 21

14 TMHELs current structure with studied system . . . 23

15 TMHEL new distribution structure with studied system . . . 24

16 Possible alternative route for supplier A . . . 25

17 Possible alternative route for supplier B . . . 26

18 Possible alternative route for supplier C . . . 27

19 Model for problem solving . . . 32

20 Schematic figure of Polya and Conway (2004) model for problem solving . . 33

21 Method structure for the thesis . . . 34

22 The duration of the thesis divided into three time phases . . . 35

23 How the problem area becomes limited by the literature gathering and finally result in a problem definition (Patel and Davidsson, 2011) . . . 35

24 The structure of how the questions were answered . . . 37

25 The transportation structure in this thesis . . . 39

26 Distribution of demand and distance to warehouses for Supplier A . . . 47

27 Total cost for the different distribution structures for Supplier A . . . 53

28 The total cost for supplier A if all warehouses had the same dynamic factor . 55 29 The total cost for supplier A if the dynamic factor in Mjölby was set to zero while the others kept Mjölbys . . . 56

30 The total cost for supplier A if the transports to Mjölby are not together with production . . . 57

31 Distribution of demand and distance to warehouses for Supplier B . . . 59

32 Total cost for the different distribution structures for Supplier B . . . 64

33 The total cost for supplier B if all warehouses had the same dynamic factor . 66 34 The total cost for supplier B if the dynamic factor in Mjölby decreased while the others used their existing dynamic factor . . . 67

35 The total cost for supplier B if freight tariffs are increased with 20% . . . 68

36 Distribution of demand and distance to warehouses for Supplier C . . . 69

37 Annual total cost for the different distribution structures for Supplier C . . . 74 38 The total cost for supplier C if all warehouses had the same dynamic factor . 76

(8)

39 The total cost for supplier C if the dynamic factor in Mjölby was set to zero while the others kept Mjölbys . . . 77 40 The total cost for supplier C if the transports to Mjölby are not together with

production . . . 78 41 The total cost for supplier C if the transports to Mjölby are not made with

(9)

List of Tables

1 Different total cost analysis models and which cost categories each include (Aronsson, 2002) . . . 16 2 Values on safety factor, k, depending on service level . . . 18 3 Summary of the main questions and sub questions for each supplier . . . 31 4 Annual transportation costs in the existing distribution structure for Supplier

A. The bold column is the inbound delivery point . . . 48 5 Annual transportation costs for the five alternative distribution structures for

Supplier A. The bold column is the inbound delivery point . . . 49 6 Annual inventory carrying costs in the existing distribution structure for

Sup-plier A. The bold column is the inbound delivery point . . . 49 7 Annual inventory carrying costs for the five alternative distribution structures

for Supplier A. The bold column is the inbound delivery point . . . 50 8 Annual warehousing costs in the existing distribution structure for Supplier

A. The bold column is the inbound delivery point . . . 51 9 Annual warehousing costs for the five alternative distribution structures for

Supplier A. The bold column is the inbound delivery point . . . 52 10 How much of the demand that needs to be moved from Mjölby to another

warehouse in order to change the choice of inbound delivery point for Supplier A . . . 54 11 Annual transportation costs in the existing distribution structure for Supplier

B. The bold column is the inbound delivery point . . . 60 12 Annual transportation costs for the five alternative distribution structures for

Supplier B. The bold column is the inbound delivery point . . . 61 13 Annual inventory carrying costs in the existing distribution structure for

Sup-plier B. The bold column is the inbound delivery point . . . 61 14 Annual inventory carrying costs for the five alternative distribution structures

for Supplier B. The bold column is the inbound delivery point . . . 62 15 Annual warehousing costs in the existing distribution structure for Supplier

B. The bold column is the inbound delivery point . . . 63 16 Annual warehousing costs for the five alternative distribution structures for

Supplier B. The bold column is the inbound delivery point . . . 63 17 How much of the demand that needs to be moved from Mjölby to another

warehouse in order to change the choice of inbound delivery point for Supplier B . . . 65 18 Annual transportation costs in the existing distribution structure for Supplier

C. The bold column is the inbound delivery point . . . 70 19 Annual transportation costs for the five alternative distribution structures for

Supplier C. The bold column is the inbound delivery point . . . 71 20 Annual inventory carrying costs in the existing distribution structure for

Sup-plier C. The bold column is the inbound delivery point . . . 71 21 Annual inventory carrying costs for the five alternative distribution structures

(10)

22 Annual warehousing costs in the existing distribution structure for Supplier C. The bold column is the inbound delivery point . . . 73 23 Annual warehousing costs for the five alternative distribution structures for

Supplier C. The bold column is the inbound delivery point . . . 73 24 How much of the demand that needs to be moved from Mjölby to another

warehouse in order to change the choice of inbound delivery point for Supplier C . . . 75

(11)

1

Introduction

This chapter contains a background description, the purpose of the study and company directi-ves.

1.1

Background

Toyota Material Handling Europe, TMHE, is one of the worlds leading forklift manufactu-rers. TMHE Logistics, TMHEL, which is a part of TMHE, are managing and storing spare parts in order to supply service technicians all over Europe (Öberg, 2018). When a forklift is sold or rented, a service deal is often established. This service deal guarantees the possi-bility of customers purchasing service and spare parts. To provide good service towards the customers to a reasonable cost, there is a need for a well functioning and fast responding supply chain (Oskarsson et al., 2013).

The structure of a supply chain is different depending on the characteristics of the product (Chopra and Meindl, 2004). According to Hu et al. (2017) there are four typical characte-ristics that spare parts have, which are; low and varying demand, large number of varieties, high risk of obsolescence and the urgent need. The goal when managing spare parts is to maximize the availability and service performance, while decreasing the economic costs by decreasing the stock levels (Fortuin and Martin, 1999). However, the difficulty in spare part management, due to the characteristics, is to decide the size of inventory, transportation, amount of needed facilities and how the information is distributed (Chopra and Meindl, 2004). If the distribution structure then changes, it affects multiple parts of the supply chain including the inventory levels and the transportation routes (Oskarsson et al., 2013). Accor-ding to Oskarsson et al. (2013) it is therefore important to have a wide scope and to study all costs when making a logistical decision.

The existing distribution structure at TMHEL is structured so that all deliveries from suppli-ers are managed by TMHELs central warehouses, CWs, before they transported to a regional or a national warehouse. That means that if a supplier is located in Germany, the products are first transported to the CW in Sweden and then transported to, for example, a regional warehouse, RW, in Germany. This leads to longer transportation routes of the spare parts, which in turn might lead to higher costs. TMHEL thereby wants to allow the supplier to deliver directly to a CW or a RW, which means that the possible inbound delivery points from the supplier increases from three warehouses to six. TMHEL then wants to investigate if there is a more cost efficient inbound delivery point from the suppliers than the existing one. (Öberg, 2018; Eriksson Bergman, 2018)

(12)

1.2

Purpose

The purpose of the thesis is to investigate if a change in inbound delivery point from three suppliers of spare parts can reduce the total cost.

Two distribution structures for spare parts will be analysed and compared. The first distri-bution structure is the current while the second is a new distridistri-bution structure. In the new distribution structure, five alternative inbound delivery points are analysed and compared with the current inbound delivery point with the current distribution structure.

1.3

Company Directives

In this thesis were five directives from the company given. The first directive was that the re-sult of the thesis should form a basis for deciding if it is cost efficient to change the inbound delivery point for three chosen suppliers. These three suppliers were chosen by TMHEL because of their different characteristics.

The second directive was that the existing service level needs to be maintained at each wa-rehouse if a change in inbound delivery point is made. The service level in this thesis refers to the availability of a product in a warehouse.

The third directive was that the inbound delivery point can only be within the TMHEL organisation, since moving goods outside the company requires a money transaction. When moving products from one TMHEL inventory from to another, no money transaction outside TMHEL is needed.

The fourth directive was that the national warehouses, distributors and service technician should be supplied by the same warehouse as before even after a change of inbound delivery point.

The last directive was that a supplier can only have one inbound delivery point, since the ordering system at TMHEL today is unable to manage that.

(13)

2

Present Situation

This chapter contains a description of the present state regarding the TMHEL organisation, the present and the new distribution structure and the three investigated suppliers.

2.1

Toyota Material Handling Europe Logistics

Toyota Material Handling Europe, TMHE is one of five divisions included in Toyota Materi-al Handling. The five divisions are Materi-all based on their geographicMateri-al market location; Europe, International, Japan, China and North America. TMH is part of the Toyota Industries Cor-poration, which consist of different companies including Toyota Material Handling, Toyota Motor Corporation and Toyota Textile Machinery. (TMHE, 2017a)

Toyota Material Handling Europe, TMHE, consist of Marketing and Sales, Rental, Service & Logistics, Logistics Solutions and Supply. The spare part management is included in Rental, Service & Logistics which also are known as TMHEL, see Figure 1. (TMHE, 2017a)

Figure 1: THME Organisation

TMHE is one of the worlds leading forklift manufacturers and consists of three major forklift manufacturers merged together, BT, Toyota and Cesab. BT and Toyota are both world lea-ding in their respective markets, BT for warehouse forklifts and Toyota for counterbalanced forklifts. TMHE has three production sites, one for warehouse forklifts located in Mjölby, Sweden, and two for counterbalanced forklifts located in Ancenis, France and in Bologna, Italy, see Figure 2. Each production site has a central warehouse, CW, which keeps an inven-tory for the spare parts. The production of warehouse forklifts only occur in Mjölby, which means that the warehouse in Mjölby works as a CW for spare parts to warehouse forklifts all over the world, not only for Europe. In the same way are the warehouses in Bologna and Ancenis also functioning as CWs for spare parts to counterbalanced forklifts all over the world. (TMHE, 2017b)

(14)

According to Eklund (2018) the CW in Mjölby, Sweden, keeps a total of 34 000 different spare parts in stock ready for immediate delivery within 24 hours. While the warehouse in Ancenis, France holds 21 000 spare parts (TMHEL, 2017a) and Bologna, Italy stores 15 000 spare parts, which makes the CW in Mjölby the largest in Europe (TMHEL, 2017b).

Additional to the CWs there are also three regional warehouses, RW, located in Antwerp, Hanover and Vienna and several TMHE national warehouses, NW, see Figure 2. TMHEL are managing all the CWs and the RWs, while the NWs are managed by TMH Market & Sales companies in their respective country. The distributors are also included in Figure 2 since they are supplied by the CWs, but they are not included in the TMHE organisation. It means that neither the NW or the distributors can be an alternative inbound delivery point in this study. Only a CW or a RW can be a potential inbound delivery point, since they are a part of TMHEL. (Eriksson Bergman, 2018)

(15)

2.2

Present Distribution Structure

TMHELs present distribution structure is structured so that all suppliers deliver directly to any of the three CWs, see Figure 3. The CWs then distributes the spare parts to all RWs, NWs and distributors. The RWs and NWs then provide the spare parts to service technicians and customers within a specific region or country. It means that the RWs the NWs and the distributors are not supplied directly by suppliers.

The CWs in Mjölby and Bologna are also functioning as a RW for a geographical area. The CW in Mjölby is a RW for Sweden, Norway and Denmark and the CW in Bologna for Italy. Worth mentioning is that the warehouse in Ancenis, France, recently became a part of TM-HEL which means that there is no current inventory level or transportation routes to and from the CW in Ancenis. The RW in Vienna is also different from the other warehouses, since they are renting the warehouse area and staff from Toyota Motor Cooperation. The CWs and RWs are currently the only warehouses that are part of TMHEL, but the goal is to include the NWs to TMHEL in the future. It means, as mentioned earlier, that only the CWs and RWs are alternative inbound delivery points. (Öberg, 2018)

Figure 3: TMHEL current structure

The main reason behind this structure is that many suppliers are the same for spare parts and for the production sites, which means that the transportation can be made together. However, the main problem with this structure is, for example, when the RW in Hanover requests spare parts from a supplier that is located near Hanover. Since the warehouse in Hanover is not a CW, the spare parts first needs to be transported to any of the three CWs. After the spare parts has been transported to the CW, they can be transported back to the RW in Hanover. (Eriksson Bergman, 2018)

(16)

To solve this problem, a reconstruction of the distribution structure has been suggested by TMHEL, see Figure 4. Since all RWs and CWs today are managed by the TMHEL organisation, it is possible for the suppliers to deliver directly to either a CW or a RW. This distribution structure gives the suppliers six, instead of three alternative inbound delivery points. (Eriksson Bergman, 2018)

Figure 4: TMHEL example on new distribution structure

2.3

Suppliers

TMHEL has many different suppliers and since the three production sites produces different kind of forklifts, some suppliers to the CWs are the same while other only supply one CW. In this thesis is the inbound delivery point for three suppliers to TMHEL investigated. The suppliers were chosen by TMHEL because of their different characteristics with either bulky, large quantities or expensive products. (Öberg, 2018)

2.3.1 Supplier A

Supplier A is located outside Nürnberg, Germany and supplies the CW in Mjölby with seats and spare parts related to seats to fork lifts. Transportation from supplier A to Mjölby is often coordinated together with the production site (Eriksson Bergman, 2018). In Figure 5, is the present structure of the transportation flow illustrated from supplier A to the CW and from the CW to the RWs. (Öberg, 2018)

(17)

Figure 5: Map of the transportation routes from supplier A

2.3.2 Supplier B

Supplier B is located in Hamburg, Germany and supplies the CW in Mjölby with different sizes of wheels to the forklift. This type of spare parts are purchased in large quantities, which requires large areas when kept in stock. (Öberg, 2018)

The transportation route to Mjöbly from the supplier, is made in two steps. The first step is that the spare parts, which are referred as items from now on, are transported from supplier B to the RW in Hanover. The second step is that the items are cross-docked in Hanover and then transported to Mjölby. Worth mentioning is that the demand for items from supplier B to the RW in Hanover, is not collected when the items are cross-docked in Hanover. The items to Hanover is first transported to the CW in Mjölby and then transported back to Hanover. In Figure 6, is the present transportation flow from Supplier B to the CW in Mjölby, and from the CW to RWs shown. (Öberg, 2018)

(18)

Figure 6: Map of the transportation routes from supplier B

2.3.3 Supplier C

Supplier C is also located outside Nürnberg, Germany and is the largest supplier of engines that TMHE have. Engines are expensive, heavy and critical items since if an engine does not work, the fork lift is standing still. A forklift that is standing still can lead to an entire assembly line to stand still, which costs a lot of money. (Öberg, 2018)

Supplier C delivers to the CW in Mjölby and the transportation from supplier C to Mjölby is made together with the production site. In Figure 7 is the present transportation flow from Supplier C to CWs and RWs shown. (Öberg, 2018)

(19)
(20)

3

Frame of References

This chapter provides the reader a more detailed background to the concepts used in this thesis which are spare parts, modelling the supply chain, total cost analysis, localisation methods and sensitivity analysis.

The structure of this chapter can be viewed as three steps connected to the purpose, ”The purpose of the thesis is to investigate if a change in inbound delivery point from three sup-pliers of spare parts can reduce the total cost”, see Figure 8.

The first step connects to the sentence ...” from three suppliers of spare parts”... and is sup-posed to introduce the reader to the theory concerning spare parts, since they are different from products regarding their lifecycle, characteristics and management (Hu et al., 2017). The second step is modelling the supply chain and connects to the sentence ...” change in inbound delivery point”..., because depending on the characteristics of the product, the supply chain can be modelled in different ways (Chopra and Meindl, 2004). If the inbound delivery point would change, the distribution structure of the supply chain would changes as well (Lumsden, 2012). It was therefore important to investigate how a supply chain of spare parts should be modelled and which parameters that are affected in a change of inbound delivery point.

The third step shows the different tools that can be used to select the most profitable in-bound delivery point and connects to the sentence ...” investigate... reduce the total cost". The first tool is a total cost analysis, which summarizes the costs connected to logistic ac-tivities (Oskarsson et al., 2013). The second tool are two location methods that determines which warehouse to use as inbound delivery point (Erlebacher and Meller, 2010; Daskin et al., 2012). The last tool is sensitivity analysis which is used to see how a result is influenced by the uncertainties in the input (Saltelli et al., 2008).

(21)

3.1

Spare Parts

Due to a longer lifecycle and varying demand, supply chain activities for products differs from spare parts regarding distribution and managing. According to Fortuin and Martin (1999) can the service period of a product be divided into three phases; the initial phase, normal phase and final phase. The lifecycle for products and the spare parts are similar to each other until the final phase, see Figure 9 (Hu et al., 2017). Spare parts have a longer lifecycle than regular products because the demand for spare parts continues even after the end of the product lifecycle (Kim et al., 2016).

The initial phase is conncented to the part of the life cycle when new parts or components are being introduced to the market (Fortuin and Martin, 1999). According to Fortuin and Martin (1999) the normal phase concerns the time when the parts are used in production and during the final phase is when production has stopped, but the service period goes on. Olhager (2013); Jonsson and Mattsson (2016) has a more detailed view of the product life-cycle and divides it into four phases; introduction phase, growth phase, maturity phase and decline phase, see Figure 9.

Figure 9: Product lifecycle and its phases, own version combining Olhager (2013) and Kim et al. (2016)

The introduction phase is similar to the initial phase described by Fortuin and Martin (1999). During this phase, it is important to be flexible at the same time maintain a high quality (Olhager, 2013). The growth phase is when the market discover the product and the demand growths (Olhager, 2013). According to Jonsson and Mattsson (2016) it is important to have short and reliable delivery times in this phase. The maturity phase is, according to Olhager (2013), the phase when the sales volume is high, the growth begins to decline and the demand begins to be fulfilled. When the demand subsequently begins to fall, the product approaches the decline phase (Olhager, 2013). In the decline phase sales reduces and the product is approaching the end of its lifecycle (Olhager, 2013). According to (Hu et al., 2017) the difficulties for managing spare parts begins when the product is approaching the end, since

(22)

3.1.1 Spare Part Characteristics

According to Hu et al. (2017), there are four typical characteristics of spare parts. The first characteristic is the varying and low demand (Fortuin and Martin, 1999; Topan et al., 2017). Due to the low demand, it becomes more important to keep the inventory carrying cost as low as possible while the effect of the ordering costs become less important, especially for expensive parts (Topan et al., 2017).

The second characteristic is the number of varieties that exists for spare parts. That is becau-se both new and ten year old spare parts needs to be available to the customers. This leads to difficulties in identifying the appropriate stock control strategy, and thereby creating a challenge for inventory management. (Hu et al., 2017)

The third one is according to Hu et al. (2017) the risk of obsolescence. In order to reduce the risk, is it important to minimize the stocks (Hu et al., 2017). However, according to Fortuin and Martin (1999) a smaller stock increases the risk of stock outs. Stock outs for spare parts can have serious consequences for the system, since a complete stop in an entire assembly line is a possible outcome (Topan et al., 2017). This leads to the necessity to keep spare parts in stock in order to compensate for long lead-times (Fortuin and Martin, 1999). The difficulty, according to Fortuin and Martin (1999) lies in the management of both the risk of obsolescence and the risk of stock outs.

The last characteristic is the urgent need of a spare part when the corresponding part of the equipment fails, is damaged or wears out. If the spare part is critical for the equipment or the whole system, it makes it even more challenging to manage. (Hu et al., 2017)

3.1.2 Managing Spare Parts

According to (Hu et al., 2017) the decision concerning which spare parts to stock and when to order them, are the most important decisions in spare parts management. When mana-ging spare parts the aim, according to Fortuin and Martin (1999); Hu et al. (2017), is to maximize the availability and service performance, while decreasing the economic costs by decreasing the stock levels. In order to maximize the availability, performance indicators such as fill rate and service rate needed are important factors, since they measure the frequency of occurrence in which clients have to wait for a particular item (Fortuin and Martin, 1999). According to Hu et al. (2017) another way to maximize the availability is to use optimization techniques, since optimization can balance capital investment and service level constraints over a large assortment of spare parts.

According to Topan et al. (2017) a high service level and availability are most important for the critical products. Because a deficient critical part can stop an entire assembly line (Topan et al., 2017). Since spare parts have a low and varying demand, it is difficult to forecast future demand (Fortuin and Martin, 1999; Hu et al., 2017). Uncertainty in a forecast is therefore often compensated by holding extra parts in inventory to increase the availability (Hu et al.,

(23)

Fortuin and Martin (1999) also states that high safety stocks may be needed to keep service rates within acceptable limits if the criticality of a part is large. However, keeping extra large stock is often unacceptable, especially for expensive spare parts since the inventory carry-ing cost increases (Fortuin and Martin, 1999). Fortuin and Martin (1999) therefore states that spare parts which have a low demand and are expensive, often has a limited availability. The costs for spare parts are often the sum of inventory carrying costs, stock-out penalty costs and ordering costs (Hu et al., 2017). According to Hu et al. (2017) is the first decision in order to decrease stock levels to decide which products to keep in stock. In order to make this decision, a categorization or classification of the service parts first needs to be done (Fortuin and Martin, 1999; Hu et al., 2017; Teixeiraa et al., 2017). According to Hu et al. (2017) there are several different criteria that a classification can be based on. The most common are however lead time and criticality, followed by annual cost and unit price (Hu et al., 2017). Fortuin and Martin (1999) also suggests that the categorization can be based on an additional criteria as demand and planning horizon.

The classification can, according to Hu et al. (2017), be made by an ABC- classification based on selected criterion. Teixeiraa et al. (2017) states another quantitative classification method called FSN, which is based on the demand patterns of the items. FSN classifies the items into three categories: fast moving, F, slow moving, S, and non-moving N (Teixeiraa et al., 2017). The slow moving items contributes, according to Syntetosa et al. (2012), little to the sales but they account for up to 60 percent of the total stock value. Therefore it is especially important to make improvements on the management of the slow moving items (Syntetosa et al., 2012).

3.2

Modelling the Supply Chain

When the supply chain is modelled, distribution needs to be taken into consideration (Jons-son and Matts(Jons-son, 2016). According to Jons(Jons-son and Matts(Jons-son (2016), distribution explains how a product is moved throughout the supply chain and where it is stored. Depending on characteristics of the product, for example high demand, high value or low desired response time, different set-ups can be used (Chopra and Meindl, 2004). If a product has a short desired response time, the supply chain requires a lower number of facilities to handle the shorter lead times (Chopra and Meindl, 2004).

The structure of the supply chain depends on total cost, customer service and the relationship between them (Chopra and Meindl, 2004). Mattsson (2012) states that a higher customer service generates a higher cost but not necessary a higher efficiency. Thereby, it is important to find the right level of customer service for each product, see Figure 10 (Mattsson, 2012).

(24)

Figure 10: Relationship between total cost and customer service, own version of Oskarsson et al. (2013)

Customer service is, according to Jonsson and Mattsson (2016), about satisfying the custo-mers needs regarding the deliveries. Before the delivery, it is important to have an understan-ding of how the delivery will be executed (Jonsson and Mattsson, 2016). During the delivery it is important to keep the agreed delivery terms and after the delivery supply with spare parts if needed (Oskarsson et al., 2013).

To measure the customer service, Lumsden (2012) lists six elements; lead time, on-time deli-very, delivery dependability, availability, flexibility and information. Lead time is, according to Lumsden (2012), the time from order to received delivery and consists of receiving order, order processing, planning, manufacturing and distribution. Oskarsson et al. (2013) says that on-time delivery is how often the deliveries are made at the agreed time. This element has gained importance lately since most companies have reduced their inventories leaving them more vulnerable for both too early and too late deliveries (Lumsden, 2012). According to Jonsson and Mattsson (2016) the delivery dependability is the ability to deliver the right product, in the right quality and in the right quantity. The importance of high delivery dependability has also increased following the changes in warehouse management (Jonsson and Mattsson, 2016). Availability describes that if the product is in inventory and is ready for immediate delivery (Lumsden, 2012). Jonsson and Mattsson (2016) says that flexibility is the ability to adjust to special demands concerning order sizes, packaging and delivery point. Information is how supplier and customer share information that can simplify the deliveries and make them more effective (Oskarsson et al., 2013).

According to Jonsson and Mattsson (2016), the aim of modelling the supply chain is to meet customers demand of service at the lowest cost as possible. If the customers do not want to pay for the highest possible service level there is no reason to dimension the supply chain for it (Jonsson and Mattsson, 2016). According to Chopra and Meindl (2004) there are four aspects that needs to be analysed when dimensioning the supply chain:

(25)

• What will the size of the inventories be?

• How is the transportation going to be carried out? • How many facilities are there going to be?

• How is the information going to be distributed?

All these four aspects are connected and a change in one of them will affect all the other (Chopra and Meindl, 2004). For example, with fewer facilities the inventory carrying cost and facility cost will decrease but the transportation cost will increase, see Figure 11 (Chopra and Meindl, 2004). Oskarsson et al. (2013) also states that if these aspects are managed in a correct way, it may be possible to reach economics of scale. According to Westin et al. (2016), there is a trend to use larger vehicles so that larger quantities can be transported to reach economies of scale. When modelling freight transport, it is therefore important to capture the impact of increased or decreased demand on the trucks Westin et al. (2016). The demand affects, according to Westin et al. (2016), the fill rates in the trucks which also affects the logistics costs, transport costs, order costs and inventory costs. The trade-off between transport costs, inventory costs and order costs thereby needs to be compromised Westin et al. (2016).

Figure 11: How costs vary depending on the number of facilities in the supply chain, own version of Chopra and Meindl (2004)

3.3

Total Cost Analysis

Oskarsson et al. (2013) and Jonsson and Mattsson (2016) states that it is important to have a wide scope and to study all costs when making a logistic decision, since a change can have an impact that decreases costs in one area and at the same time increases costs in another. The costs are often divided into different categories connected to different parts of the supply chain (Waller and Fawcett, 2012). Aronsson (2002) says that more specific cost categories can reduce the risk for multiple calculation. Which these different cost categories should be, have been debated and different models have been presented (Waller and Fawcett, 2012).

(26)

Aronsson (2002) has investigated several models with different cost categories, and the most common categories included in the model were; inventory carrying cost, warehousing cost and transportation cost, see Table 1. In addition to that, most of the models included a category which took the administration cost in consideration (Aronsson, 2002). In which cost category a specific cost is placed, is not so important since all costs are summarized and will therefore affect the total cost (Oskarsson et al., 2013).

Table 1: Different total cost analysis models and which cost categories each include (Aronsson, 2002)

Model Abrahamsson & Aronsson (1999) Co yle et al. (199 2) Dagzano (1999) P ersson & Virum (1996) Lewis & Cullingtion (1956) Delaney (2000) Harrison (1999) F riedman (1997) HRF OCUS: Drey er (2000) Alfak onsult (1996) Christopher (1998) Transportation cost x x x x x x x x x x x Administration cost x x x x x

Inventory carrying cost x x x x x x x x x x

Warehousing cost x x x x x x x x x x

Packaging cost x

Cost for return of goods x

Cost of missing sales x x x

Communication and IT x

Production cost x

Order processing cost x x x x

Ordering cost x x

Other costs x

Oskarsson et al. (2013) presents a similar total cost model, based on the model by Lambert (1976), which can be adjusted to fit each unique situation. The investigated cost categories for the model by Oskarsson et al. (2013) are inventory carrying cost, warehousing cost, transportation cost, administration cost and other costs. The framework for this thesis will consist of the cost categories presented by Oskarsson et al. (2013), since many of the models which Aronsson (2002) studied also had a cost category similar to Oskarsson et al. (2013) category ”other costs” and more detailed categories reduce the risk for multiple calculation. Then to get the total cost, all cost categories from the model by Oskarsson et al. (2013) needs to be summarized, see Equation 1 (Oskarsson et al., 2013).

Total cost, TC =XInventory carrying cost + Warehousing cost+

(27)

3.3.1 Inventory Carrying Cost

Inventory carrying cost is, according to Jonsson and Mattsson (2016), the cost for keeping products in stock, which can be divided into risk cost and cost of tied up capital. The risk cost, is a costs for obsolescence, insurance and pilferage (Lambert, 1976). The tied up capi-tal cost is, according to Jonsson and Mattsson (2016), the cost that occur when the capicapi-tal is frozen in the goods and it is not possible to invest the capital to receive return on an investment. The cost of tied up capital depends on average stock level and if the stock le-vels increases the costs will increase (Oskarsson et al., 2013). To understand how the cost varies over time, the average stock level needs to be further described (Oskarsson et al., 2013). The average stock level consists of the cycle stock, CS, and the safety stock, SS, see Equation 2. The average cycle stock, see Equation 3, is half the order quantity. The order quantity is in turn based on the economical order quantity, see Equation5. The safety stock, SS, is the stock used to prevent stock outs and it depends on uncertainties in demand and lead time, see Equation 4 and Table 2. The safety stock was in this thesis was calculated according to SERV1, see Table 2, which is based on the possibility not to have a shortage during a stock cycle. The inventory carrying cost is equal to the average stock level multiplied with the value of the product and an interest rate, see Equation 6. (Oskarsson et al., 2013)

Average stock level, ASL = CS + SS = Q 2 + k q σ2 DLT + σ2LTD2 (2) Cycle stock, CS = Q 2 (3) Safety stock, SS = kqσ2 DLT + σLT2 D2 (4)

Economical Order Quantity, EOQ =

s

2KD

rp (5)

Inventory carrying cost, ICC = ASLpr =

Q 2 + k q σ2 DLT + σLT2 D2  pr (6)

K = Fixed order cost

D = Demand

r = Interest rate p = Product value k = Safety factor LT = Lead time

σD = Standard deviation in demand

σLT = Standard deviation in lead time

(28)

Table 2: Values on safety factor, k, depending on service level

Service level, (%) 50 90 95 98 99 99,5 Safety factor, k 0,00 1,28 1,64 2,05 2,33 2,58

To calculate the cycle stock by using Equation 3 and 5, the demand needs to be stable (Oskarsson et al., 2013). If the demand varies, it is better to use measured values to calculate the cycle stock (Oskarsson et al., 2013). A varying demand might also require a higher safety stock, according to Oskarsson et al. (2013), since the order quantity also varies. When the order quantity varies it can pass the ordering-point, see Figure 12 (Oskarsson et al., 2013). Passing the ordering point can create less items in stock than calculated, which might lead to stock outs (Mattsson, 2005).

Figure 12: Ordering-point system with varying demand, own version of Mattsson (2005)

3.3.2 Warehousing Cost

The warehousing cost consists, according to Lambert (1976), consists of the costs for having the warehouse. It includes the costs for the building, staff, equipment used in the warehouse and the movement of goods inside the facility, see equation 7 (Jonsson and Mattsson, 2016). According to Oskarsson et al. (2013) will a change in order quantity affect the frequency of the deliveries. A decrease in order quantity may lead to more frequent deliveries if the demand is kept constant (Oskarsson et al., 2013). More frequent deliveries might thereby require more staff and equipment, which would increase the warehousing cost (Oskarsson et al., 2013).

Warehousing cost, W =XBuilding cost, Staff cost,

(29)

3.3.3 Transportation Cost

According to Lambert (1976) and Oskarsson et al. (2013) is the transportation cost the costs for deliveries outside the company and includes transportation between a company’s facilities and to external companies, see equation 8. Lumsden (2012) divides transportation costs in the two groups; costs that often are connected with the movement of goods, and costs that are not connected with the movement. Regular costs which are connected with the movement of goods are loading, transshipment and unloading (Lumsden, 2012). Oskarsson et al. (2013) says however that only transshipment is connected with the movement of goods and that loading and unloading are warehousing costs. Costs that are not connected with the movement can, according to Lumsden (2012), be inventory holding cost during transpor-tation, cost for damaged goods, insurance costs, costs to ensure that a third party logistics company carries out the transport and administration costs for planning a transport.

Transportation cost, T =XCosts between a companies facilities and

costs to external companies (8)

3.3.4 Administration Cost

According to Jonsson and Mattsson (2016) are the administration costs the costs connected to planing and order processing. It includes wages for administrative staff and costs for computer systems, see equation 9 (Jonsson and Mattsson, 2016).

Administration cost = KD

Q (9)

K = Fixed order cost

D = Demand

Q = Actual order quantity 3.3.5 Other Costs

Other costs are logistic costs that can be difficult to place under one of the other four categories (Oskarsson et al., 2013). These could be, according to Aronsson (2002), costs for material, packaging, missing sales and returning of goods.

3.4

Localisation Methods

According to Daskin et al. (2012) and Lumsden (2012) the efficiency of the distribution sy-stem is dependent on where the goods are produced and the localisation of the terminals. Poorly planned locations of warehouses can result in excessive costs and degraded service (Daskin et al., 2012). To have an efficient supply chain, the most critical and difficult decision is therefore where to locate a terminal (Daskin et al., 2012).

(30)

When choosing a terminal location or distribution breakpoint, there are several other factors that affects the decision (Lumsden, 2012). According to Lumsden (2012) the geographical location, infrastructure and the economic conditions are some of these factors. The most common method used when choosing the location of a distribution network, is to investigate where the density of the customers need is located (Lumsden, 2012). The definition of the density analysis is thereby presented below. A location model, which is another localisation method, decides where to locate a terminal/distribution breakpoint by minimizing both transportation costs and inventory carrying costs is also described.

3.4.1 Centre of Demand Density

The purpose of the density analysis is to design the distribution network so that the wa-rehouses are where the customers need is largest (Olhager, 2013). A density analysis is, according to Lumsden (2012), usually based on the customer demand and location since the transportation work for the distribution thereby gets minimized. This density analysis also requires that the cost for transporting is equal regardless where in the area the goods are produced and how much that is transported (Lumsden, 2012).

According to Grant et al. (2006) the density analysis is thereby often used to give a first view on where to locate a central warehouse or distribution breakpoint. But in order to achieve satisfactory results it needs to be modified to take into account factors such as time and customers service level.

The method can be used for different cases of localisation, one of the cases is when a supplier distributes through a terminal to the customers (Lumsden, 2012). To find the centre of demand, each customer needs to have a predefined demand, according to Lumsden (2012). Grant et al. (2006) describes this as several pieces of rope tied together in the middle of a circular area, with different weights at the end of each rope. Initially, the knot would be located in the middle of the circle. When the weights then are released, the centre of demand would be at the point where the weights are in balance (Grant et al., 2006). Lumsden (2012) describes this in another more detailed way where each rope represent one customer and the placement around the circle is the customers x- and y coordinates. The weights at the end of each rope corresponds to the demanded quantity for each customer (Lumsden, 2012), see Figure 13. By viewing the relationship between demand and distance, the optimal localisation of the distribution breakpoint can be determined (Lumsden, 2012).

(31)

Figure 13: Density of demand analysis, Lumsden (2012)

3.4.2 Localisation Model

To make a decision on where to locate a distribution breakpoint or facility, a localisation model is often used. There are several existing localisation models but the model which forms the basis of many other localisation models is, according to Daskin et al. (2012), called The Fixed Charge Facility Location Problem or The Generalized Weber problem, according to Er-lebacher and Meller (2010). This model can be described as a set of customers with a known demand and a set of potential facility locations. The objective is to decide the number of facilities needed and were to locate them (Erlebacher and Meller, 2010). However, if the locations of the facilities are known, the problem is instead called The traditional transpor-tation problem, according to Lundgren et al. (2010) and Daskin et al. (2012). This model allows only one facility to be the distribution breakpoint and that each facility has a capacity constraint of the total amount of goods it can manage to make the model realistic (Daskin et al., 2012). The objective for this model is then to minimize the cost for the quantity ship-ped between the customer and the facility (Daskin et al., 2012). To make the model more realistic, capacity constraints of the warehouses can be added (Daskin et al., 2012)

(32)

However, a localisation decision should not be based only on the transportation cost (Da-skin et al., 2012). According to Da(Da-skin et al. (2012) the contribution of inventory costs to distribution costs has been recognized for many years. For example, when the number of facilities increase, the inventory carrying costs as well as the facility cost increases, while the transport cost decrease (Chopra and Meindl, 2004; Daskin et al., 2012). Both Nozick and Turnquist (2001) and Daskin et al. (2012) have therefore developed their own location models that minimizes both inventory and distribution costs. The model that Daskin et al. (2012) has developed, has the objective to minimize the fixed facility location costs, direct transportation costs to the customers and inventory carrying costs (Daskin et al., 2012).

3.5

Sensitivity Analysis

According to Lamboni (2018), models of physical or natural phenomena are often used as an experimental tool for supporting decision making. However, models often include several estimations and uncertainties in the input data, which might have a strong effect on the output of the model (Lamboni, 2018). A model can thereby not be validated, in the sense of being proved true, before it has survived a series of tests which shows if the model can explain or predict the ”right behaviour” in a convincing way (Saltelli et al., 2008; Oskarsson et al., 2013). The purpose of a sensitivity analysis is therefore, according to Yeung et al. (2010), to see how the network output is influenced by its input or disturbances. Saltelli et al. (2008) defines sensitivity analysis in a similar way,

”The study of how uncertainty in the output of a model can be apportioned to different sour-ces of uncertainty in the model input.”

In a sensitivity analysis, the purpose is to determine which of the input parameters that are more important in influencing the uncertainty in the model (Saltelli et al., 2008). The input parameters can, according to Oskarsson et al. (2013), either be varied one by one at the same time as all other parameters are kept constant, or they can be varied in different combinations. The input parameters that should be analysed and varied are the estimated parameters and not the mathematical constants or the internal model variables (Saltelli et al., 2008).

(33)

4

Problem Definition

This chapter describes the studied system and breaks down the purpose into two main ques-tions which will help answer the purpose for each studied supplier.

4.1

The Studied System

The studied system for the current situation was chosen to be within the borders of TMHEL and the distribution from the supplier, see Figure 14. The NWs and distributors were not included in the studied system, since they are not part of TMHEL. The transportation from CW to NWs and distributors were also excluded since the sales companies connected to each NW or distributor are responsible for the transportation and the corresponding costs. The service technicians were also excluded from the system since a change of inbound delivery point would not affect them.

Figure 14: TMHELs current structure with studied system

For the new distribution structure the studied system is the same, which makes it possible to compare the costs for the current and the new structure, see Figure 15. The five alterna-tive inbound delivery points investigated are thereby the warehouses in Antwerp, Bologna, Ancenis, Hanover and Vienna.

(34)

Figure 15: TMHEL new distribution structure with studied system

4.1.1 Alternative Situation

To visualize which alternative inbound delivery point that can occur and how the transpor-tation routes would change, three examples for supplier A, B and C are shown in Figure 16, 17 and 18 below. The three examples are based on selecting the warehouse closest to the supplier to reduce the transportation cost. However, these are just examples, all warehouses needs to be investigated before it is possible to say which warehouse that is the most cost efficient to use as the inbound delivery point.

For supplier A is the RW in Hanover an alternative inbound delivery point, see Figure 16. It means that if the RW in Hanover would be the inbound delivery point, Hanover is responsible to store and distribute all items to all other CWs and RWs. Using Hanover as the inbound delivery point would reduce the transportation distance from the supplier to the inbound delivery point.

(35)

Figure 16: Possible alternative route for supplier A

An alternative solution for supplier B be to use the RW in Hanover as the new inbound delivery point, see Figure 17. This alternative solution would also decrease the transportation distances from the supplier to the inbound delivery point.

(36)

Figure 17: Possible alternative route for supplier B

For supplier C an alternative inbound delivery point could also be the RW in Hanover, since it would decrease the transportation distance, see Figure 18. The RW in Vienna could also be an alternative solution, since the transportation distance from supplier C to Vienna and supplier C to Hanover is almost the same.

(37)

Figure 18: Possible alternative route for supplier C

If these three example solutions for supplier A, B and C gives a lower total cost than the current inbound delivery point, is not known. However, it would probably lead to lower transportation costs, since the transportation distances becomes shorter. It is also not known if the RW in Germany has the capacity to store and distribute to all the other CWs and RWs, as the CW in Mjölby have. If the RW in Germany does not have the capacity to store all items, it can be more profitable to distribute the items from another RW or CW.

4.2

Purpose Breakdown

”The purpose of the thesis is to investigate if a change in inbound delivery point from three suppliers of spare parts, can reduce the total cost”

In order to answer the purpose, it was divided into two main questions A and B. Main question A aims to investigate the total cost with the current inbound delivery point and the alternative distribution structure with five other possible inbound delivery points. Main question B aims to analyse the results in main question A through a sensitivity analysis.

(38)

4.2.1 Main Question A

The distribution structure today leads to long transportation routes and lead times, which can lead to higher transportation costs and inventory carrying costs. It is therefore neces-sary to investigate what the current costs are and if another inbound delivery point could reduce the total cost. The total cost therefore needs to be calculated for the current inbound delivery point and the five alternative inbound delivery points.

The total cost can be calculated in different ways, for example with a total cost analysis or with another localisation method, as described in Chapter 3.3 and 3.4. According to Arons-son (2002) the total cost is equal to the sum of different costs, which often are the cost for transportation, inventory, warehouse and administration. The decision of which costs to include in the total cost, needs be taken into consideration for each specific case (Oskarsson et al., 2013). In this thesis is the total cost model by Oskarsson et al. (2013) used, which involves the transportation cost, inventory carrying cost, warehousing cost, administration cost and other costs. The administration cost is excluded, since a change of inbound delivery point would solely move the administration cost from one warehouse to another while being kept constant in regards of total cost (Öberg, 2018; Eriksson Bergman, 2018). Other costs were also excluded in this thesis, since it is too difficult to estimate these parameters for a new inbound delivery point. To determine if a change of inbound delivery point could reduce the total cost, the first main question is formulated as:

A Which inbound delivery point gives the lowest total cost for each supplier?

To answer question A, it is divided further into six sub questions. To investigate if a more cost efficient inbound delivery point than the current exists, the current total cost based on the transportation costs, inventory carrying costs and warehousing cost needs to be calculated. The total cost for the five alternative inbound delivery points also needs to be calculated, in order to compare to the current total cost.

When the inbound delivery point changes, the distribution structure changes, which leads to new transportation routes (Chopra and Meindl, 2004). When the transportation routes changes, it leads to new transportation costs. The current and new transportation costs therefore needs to be analysed and compared, thereby sub question A1 and A2 are formulated as follows:

A1 What is the existing transportation cost?

(39)

A change of transportation routes also affect the lead time (Lumsden, 2012). According to Lumsden (2012) a change in lead time may affect the service level, safety stock and the average inventory level, which in turn affects the inventory carrying cost. The order quantity for the current and the new inbound delivery point will also be affected in a change of inbound delivery point (Jonsson and Mattsson, 2016). If the order quantity changes, it also affects the inventory levels and the corresponding costs, according to Jonsson and Mattsson (2016), see equation 2 Chapter 3.3.1. The new inventory levels thereby needs to be calculated and analysed. Sub questions A3 and A4 are thereby formulated as:

A3 What is the existing inventory carrying cost?

A4 What is the inventory carrying cost for the five alternative inbound delivery points?

A new order quantity may result in a change of the delivery frequency, according to Jonsson and Mattsson (2016). Another delivery frequency affects how much storage space, staff and equipment needed in the warehouse, which results in new warehousing costs (Oskarsson et al., 2013). Since a change of inbound delivery point can affect the delivery frequency, sub questions A5 and A6 are consequently formulated:

A5 What is the existing warehousing cost?

A6 What is the warehousing cost for the five alternative inbound delivery points? 4.2.2 Main Question B

According to Yeung et al. (2010) the purpose of a sensitivity analysis is to investigate how the network output is influenced by its input or disturbances. Since the total cost analysis in this thesis includes estimated data and the validity of the total cost model cannot be proved ”true” before it has survived a series of test, it is necessary to investigate if the result would change if changes in the data would occur (Saltelli et al., 2008). The second main question is therefore:

B How would changes in the input data affect the total cost?

According to Topan et al. (2017), spare parts have a low and varying demand. When the demand increases or decreases both the transportation cost and the inventory costs are af-fected. The transportation cost is affected, since it affects the number of pallets transported between the warehouses. The inventory carrying cost is affected, since the demand affects the order quantity. The order quantity in turn affects the cycle stock and the standard devi-ation of the demand, which affects the safety stock. The warehousing cost is also affected at the warehouses. However, since the average salary is assumed to be equal at all warehouses, it does not have an impact on the total cost. The warehousing cost only moves from one warehouse to another, while the total demand remains the same.

(40)

It is therefore necessary to investigate if the inbound delivery point, chosen when answering main question A, would change if the demand increased or decreased. Therefore is sub question B1 formulated:

B1 Would the choice of inbound delivery point change if the demand increases or decreases?

The fourth typical characteristic of spare parts is, according to Hu et al. (2017), the urgent need when the corresponding part of the equipment fails. If there is an urgent need for a critical part, it leads to the necessity to keep a larger stock of items at the warehouses (For-tuin and Martin, 1999). For spare parts that are frequently ordered, TMHE has an extra large safety stock. The extra safety stock is estimated to be different when a warehouse is the inbound delivery point and when it is not the inbound delivery point. However, the extra safety stock for Mjölby is estimated to be the same even if Mjölby is not the inbound delivery point. That is because it is not possible to estimate what the new extra safety stock would be if Mjölby was not the inbound delivery point. It is thereby necessary to investigate what the result would be, if the extra safety stock is estimated differently.

When the warehouses are not the inbound delivery point, the estimation for the extra stock is different for all warehouses. Thereby it is also interesting to investigate what the results would be if the extra safety stock was estimated to be the same for all warehouses. Sub question B2 is thereby formulated:

B2 How would the result change if the extra safety stock was estimated differently? Since the transportation from the supplier currently is made together with the production site, TMHEL is not charged by a freight tariff. In order to calculate what the transporta-tion cost would be from the supplier to the other five warehouses, a freight tariff for other items was used as an estimation. According to Saltelli et al. (2008) it is important to do a sensitivity analysis on estimated input data. It is thereby necessary to investigate what the result would be if the freight tariff changes. It is also interesting to investigate if the result changed if Mjölby is not allowed to transport together with production when being the inbound delivery point.

Transportation routes between all warehouses, does not exist today. Thereby an estimation is made for the non existing transportation routes, based on the average cost per pallet and kilometre that the existing transportation routes have. It is thereby interesting to investigate what the results would be if the estimated cost per pallet changed. Sub question B3 is thereby formulated:

(41)

4.3

Summary of the Questions

In Table 3, the two main questions and the sub questions are summarized.

Table 3: Summary of the main questions and sub questions for each supplier

A Which inbound delivery point gives the lowest total cost for each supplier? A1 What is the existing transportation cost?

A2 What is the transportation cost for the five alternative inbound delivery points? A3 What is the existing inventory carrying cost?

A4 What is the inventory carrying cost for the five alternative inbound delivery points? A5 What is the existing warehousing cost?

A6 What is the warehousing cost for the five alternative inbound delivery points? B How would changes in input data affect the total cost?

B1 Would the choice of inbound delivery point change if the demand increases or decreases?

B2 How would the result change if the extra safety stock was estimated different? B3 How would the result change if the conditions to transport changed?

(42)

5

Method

This chapter describes the choice of method in this thesis. The method is divided into three phases, understanding the problem, planning and execution and analysis. In the last part of the chapter the choice of method analysed and some critical aspects are discussed.

Björklund and Paulsson (2012) states that in projects of larger size it is important to use a predefined and general model for problem solving to ensure a high academic value of the results. Lekvall and Wahlbin (2001) continues and points at the importance to know which steps the process contains and how each step contributes to solving the problem. Lekvall and Wahlbin (2001) describes a model, which divides the process into eight steps, see Figure 19. Lekvall and Wahlbin (2001) says that the model is mostly for marketing decisions, but since the steps have general characteristics it is possible to apply on other problems as well. The first four steps aim primarily to understand the problem and to plan how to solve it, while the last four steps are about executing the plan and analysing the results (Lekvall and Wahlbin, 2001).

Figure 19: Model for problem solving, own version of (Lekvall and Wahlbin, 2001)

Polya and Conway (2004) present an additional model and describes the route to solve a problem in four phases; understanding the problem, devising a plan, carrying out a plan and look back at the problem, see Figure 20. The model by Polya and Conway (2004) is created for mathematical problems, but since the steps are relevant for problem solving in general, just like the model by Lekvall and Wahlbin (2001), it can be adjusted for other problems. In the first phase, understanding the problem, it is important to gather information and to create questions to know what to answer (Polya and Conway, 2004). Devising a plan entails answering the questions found in the problem phase (Polya and Conway, 2004). Polya and Conway (2004) continues by saying that carrying out a plan is about executing the steps developed in previous phase, and looking back is about analysing the results in order to de-velop greater knowledge.

(43)

Figure 20: Schematic figure of Polya and Conway (2004) model for problem solving

From the two described models, a model for this thesis was developed. The model used in this thesis was divided into three phases; Understanding the problem, Planning and, Execution and analysis, see Figure 21. The first phase, Understanding the problem, is similar to the first step in the model by Polya and Conway (2004) and the two first steps in the model by Lekvall and Wahlbin (2001), which is about understanding the problem. The second phase, Planning, is similar to the second step in the model by Polya and Conway (2004) and the third and fourth step in the model by Lekvall and Wahlbin (2001), which is about creating a plan for how the problem should be solved. The last phase, Execution and analysis, is a combination of the last two steps in the model by Polya and Conway (2004), which is carrying out the plan and look back at the problem, and the last four steps in the model by Lekvall and Wahlbin (2001). A more detailed description of each phase is described in the sub chapters below. At the start and between the three phases, meetings with the supervisor at the university were held. The purpose of these meetings was to get feedback on what was written so far and to make sure that the time plan had been followed.

References

Related documents

Some previous studies have already shown the risk of floor covering in room corridors set-ups and the lack of correlation with existing legislative methods both for room

It sets forth a methodological framework for estimating costs for society related to negative impacts on the environment and human health, including health- related costs and costs

46 Konkreta exempel skulle kunna vara främjandeinsatser för affärsänglar/affärsängelnätverk, skapa arenor där aktörer från utbuds- och efterfrågesidan kan mötas eller

Both Brazil and Sweden have made bilateral cooperation in areas of technology and innovation a top priority. It has been formalized in a series of agreements and made explicit

För att uppskatta den totala effekten av reformerna måste dock hänsyn tas till såväl samt- liga priseffekter som sammansättningseffekter, till följd av ökad försäljningsandel

The increasing availability of data and attention to services has increased the understanding of the contribution of services to innovation and productivity in

Electric buses have a very low operating cost compared with conventional buses, but a higher investment cost of the battery and chargers, and sometimes additional cost of the

Four main conclusions can be drawn based on the results of this thesis: BEVs can be TCO competitive with ICEVs, TCO is rarely applied in the vehicle choice process, the market