Analysis of vehicle utilization for Safcor Panalpina

(1)

Department of Science and Technology Institutionen för teknik och naturvetenskap Linköping University Linköpings Universitet

SE-601 74 Norrköping, Sweden 601 74 Norrköping

LiU-ITN-TEK-A--08/089--SE

Analysis of vehicle

utilization for Safcor

Panalpina

Paulina Engerberg

Patrik Martinsson

(2)

LiU-ITN-TEK-A--08/089--SE

Analysis of vehicle

utilization for Safcor

Panalpina

Examensarbete utfört i kommunikations- och transportsystem

vid Tekniska Högskolan vid

Linköpings universitet

Paulina Engerberg

Patrik Martinsson

Handledare Krisjanis Steins

Handledare Quintus Carstens

Examinator Clas Rydergren

Norrköping 2008-06-16

(3)

Upphovsrätt

Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare –

under en längre tid från publiceringsdatum under förutsättning att inga

extra-ordinära 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 det lösningar av teknisk och administrativ

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 considerable time from the date of publication barring

exceptional circumstances.

The online availability of the document implies a permanent permission for

anyone to read, to download, to print out single copies for your own use and to

use it unchanged for any non-commercial research and educational purpose.

Subsequent transfers of copyright cannot revoke this permission. All other uses

of the document are conditional on the consent of the copyright owner. The

publisher has taken technical and administrative measures to assure 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/

(4)

1

Master thesis

Analysis of vehicle utilization

for Safcor Panalpina

Paulina Engerberg

Patrik Martinsson

2008-06-30

Department of Science and Technology Institutionen för teknik och naturvetenskap

Linköpings Universitet Linköpings Universitet

(5)

2

Abstract

Safcor Panalpina (SAPA) is situated in South Africa and acts as a forwarder as well as a provider of supply chain solutions to clients all over the world. Due to keen competition, it is essential for Safcor Panalpina to continuously improve their business and enhance the

efficiency throughout the chain. In order to achieve a higher profit, Safcor Panalpina needs to render a more efficient organization and face the competition from other actors on the market. In the present situation, Safcor Panalpina is able to supply clients by their own vehicle fleet or by an outsourced company called MPISI, which is a Black Economical Empowerment (BEE) company. BEE companies are black-owned and are a common used term introduced by the South African government, aimed to help black ownership grow their businesses.

This thesis is mainly based on the cargo and vehicle flows from two facilities, Phase 1 and 2, to the clients in the Gauteng area, which is a province in South Africa including large cities such as Johannesburg and Pretoria. The facilities are based adjacent to the International Airport of Johannesburg.

Safcor Panalpina claims that the utilization level for all vehicles is unacceptably low in the current situation. Since SAPA intend to reduce their own fleet, in order to expand the relationship with MPISI, they want to find the right mixture of vehicles economically viable for both parties. The purpose of this thesis is consequently to examine the utilization level of the SAPA and MPISI fleets today and further on find a new optimal fleet of vehicles. Hence, the thesis will investigate the cost perspective of the issue in order to understand what cost savings that can be achieved with this new optimal fleet.

The chosen method for this thesis is an operation research approach entitled computer simulation, aiming to execute experiments, scenarios and finding the utilization level of each vehicle. The simulation model that is based on the real system, ended up to be a complex depiction of the actual system. Furthermore, experiments and scenarios were performed in order to find better solutions to the problem. However, the cost aspect has as well been taken into consideration in the simulation for all performed scenarios. This facilitates the cost comparisons between the original scenario, in other words the original simulation model, and the executed scenarios. The most cost saving alternative as well as the one with highest utilization level for the vehicles is consequently the chosen alternative in the end.

(6)

3

Acknowledgement

We would like to thank all the staff at Safcor Panalpina in the Gauteng region that we been in contact with, especially Greg James, Jacques Erasmus, Jean Staniforth, Ahmed Suleman, Winston Mullany, Lindsey Peters and our tutors Quintus Carstens and Anthonie Verploegh for their assistance. All of them have been patient with our endless questions and have given us valuable help all the way through the project. Moreover, thank you to all the helpful people at the terminals Phase 1 and 2 and at Wrench Road.

Finally, we would like to give a special thank you to our tutors at Linkoping University in Sweden, Krisjanis Steins and Clas Rydegren for their professional guidance throughout the thesis.

(7)

4

LIST OF CONTENT

1 INTRODUCTION ...11

1.1 Background ... 11

1.2 Presentation of the problem ... 11

1.3 Objectives ... 12

1.4 Delimitations ... 12

1.5 Outline ... 13

2 ABOUT SAFCOR PANALPINA ... 14

2.1 Safcor Panalpina ... 14

2.1.1 Relationship to Panalpina ... 15

2.1.2 Relationship to MPISI ... 15

3 PRESENT SITUATION ... 16

3.1 Phase 2 ... 16

3.2 Flow of cargo from airport to client ... 17

3.2.1 Dispatch at Phase 2 ... 18

3.2.2 Transportation ... 19

3.3 Safcor Panalpina and MPISI vehicle fleets... 20

4 FRAME OF REFERENCE ...21

4.1 Logistics ... 21

4.1.2 Delivery service... 21

4.1.3 Relationship between delivery service and profitability ... 22

4.2 Transportation ... 23

4.2.1 Relationship between logistics and transportation ... 23

4.3 Resource utilization of vehicles ... 25

4.3 Operations research ... 25

4.3.1 Three different approaches ... 26

4.4 Computer simulation ... 26

4.4.1 Simulation versus simulator...27

4.4.2 Considerations before using simulation ...27

4.4.3 Classification of simulation models ... 29

4.4.4 Procedure in a simulation study... 29

4.4.5 Verification and Validation ... 30

4.5 Optimization modelling ... 32

(8)

5

4.5.2 Considerations before using optimization ...35

4.6 Heuristic search ...35

4.6.1 Heuristic methods for solving route planning problems ... 36

4.6.2 Considerations before using heuristics ... 39

5 METHOD ... 40

5.1 Method of procedure ... 40

5.2 Data collection ... 42

5.2.1 Data collection in general ... 42

5.2.2 Data collection for a simulation model ... 42

5.3 Sources of errors ... 43

6 CONCEPTUAL MODEL ... 44

6.1 Overview of vehicle and cargo flow ... 44

6.2 Level of detail ... 45

6.3 Modelling assumptions ... 46

6.4 Logical flow ... 47

7 DATA COLLECTION ... 50

7.1 Data collected for simulation model ... 50

7.1.1 Data from the MySQL database ... 50

7.1.2 Time matrix ... 52

7.1.3 Weight and volumetric weight data ... 52

7.1.4 Cages ...53

7.1.5 Other modelling input data ...53

7.2 Other data collected ...53

7.2.1 Diagrams ...53

7.2.2 Cargo delivered by MPISI fleet ... 55

7.2.3 Cargo delivered by SAPA fleet ... 55

7.2.4 Modification of the data ... 56

7.2.5 Transportation costs ... 57

8 DESCRIPTION OF THE MODEL ... 61

8.1 Presentation of the simulation model... 61

8.1.1 Cargo flow in the model ... 61

8.1.2 Vehicle flow in the model... 62

8.1.3 Animation ... 68

(9)

6

9.1 Verify and validate... 69

9.1.1 Verifying the model ... 69

9.1.2 Verifying data ...70

9.1.3 Validation ...70

9.2 Presentation of scenarios ... 74

9.3 Results of scenarios ... 77

9.3.1 Original scenario ... 77

9.3.2 Scenario 1 – exchange F817 for a eight tons MPISI vehicle ... 80

9.3.3 Scenario 2 – number of vehicles needed ... 80

9.3.4 Scenario 3 – vehicle fleet reduced ... 82

9.3.5 Scenario 4 – exclude smaller deliveries ... 86

9.3.6 Scenario 5 – exclude smaller deliveries and reduce vehicle fleet ... 88

9.3.7 Scenario 6 – joint loadings between Phase 1 and 2 and reduced fleet ... 90

9.4 Scenarios from a cost perspective ... 92

9.4.1 Cost results for original scenario ... 92

9.4.2 Cost result for scenario 1 ... 93

10 RESULTS AND DISCUSSION ... 104

10.1 Future research ... 111

REFERENCES ... 112

LIST OF APPENDIXES ... 114

Appendix 1: Cages... 114

Appendix 2: Detailed information regarding the vehicle fleets... 115

Appendix 3: Time matrix ... 116

Appendix 4: Unloading time and delivery area for each client ... 118

Appendix 5: Transportation cost for F412 ... 119

Appendix 6: Utilization level from five perspectives for Safcor Panalpina vehicles ... 120

Appendix 7: Utilization level from five perspectives for MPISI vehicles ... 121

Appendix 8: Volume and time utilization from Scenario 2 ... 122

(10)

7

Appendix 10: Volume and time utilization from Scenario 3.3 ... 126

Appendix 11: Volume and time utilization from Scenario 3.5... 127

Appendix 12: Volume and time utilization from Scenario 4 ... 129

(11)

8

LIST OF FIGURES AND TABLES

Figures

Figure 2.1: Profit for each region ... 15

Figure 3.1: Phase 1 and Phase 2 location ... 17

Figure 3.2: An overview of Phase 2 and the cargo flow from Phase 2 to client ... 19

Figure 4.1: Relationship between the delivery service level and profit ... 23

Figure 4.2: The optimization procedure ... 33

Figure 4.3: Illustration of a Route Planning Problem ... 35

Figure 4.4: The basic problem for a route problem by using heuristics ... 37

Figure 4.5: Illustration of the result by using Clark & Wright ... 37

Figure 5.1: Procedure for the thesis ... 41

Figure 6.1: Logical relationship between the cargo and cages as well as vehicles and cages ... 45

Figure 6.2: Conceptual model ... 49

Figure 7.1: The actual payload and volumetric weight for each channel during the time period ... 54

Figure 7.2: The actual payload each day during September 2007 ... 54

Figure 7.3: Loaded weight for each MPISI vehicle during August and September ... 55

Figure 7.4: Loaded weight for each SAPA vehicle during August and September ... 46

Figure 7.5: The modified delivery weight for each day ... 57

Figure 8.1: The model consists of five sub models ... 61

Figure 8.2: Illustrates the vehicle creating procedure in the model ... 62

Figure 8.3: Shows the vehicle flow for one of the cages ... 63

Figure 8.4: The drop off module and the logical procedure when unloading the cargo ... 66

Figure 8.5: Illustrates the animation ... 68

Figure 9.1: The utilization level from a weight and volume perspective for the SAPA fleet ...77

Figure 9.2: The utilization level from a weight and volume perspective for the MPISI fleet ... 78

Figure 9.3: The utilization level from a time perspective for the SAPA fleet ... 78

Figure 9.4: The utilization level from a time perspective for the MPISI fleet ... 79

Figure 9.5: The physical utilization level for the SAPA fleet from the original scenario ... 79

Figure 9.6: The physical utilization level for the MPISI fleet from the original scenario ... 80

Figure 9.7: The physical utilization level for the SAPA fleet from the scenario 2 ... 81

(12)

9

Figure 9.9: The physical utilization level for the SAPA fleet from the scenario 3.1 ... 83

Figure 9.10: The physical utilization level for the MPISI fleet from the scenario 3.1 ... 83

Figure 9.15: The physical utilization level for the MPISI fleet from the scenario 4 ... 87

Figure 9.16: The physical utilization level for the SAPA fleet from the scenario 5,1 ... 88

Figure 9.20: The physical utilization level for the MPISI fleet from the scenario 6 ... 91

Figure 10.1: Safcor Panalpina’s utilization level from load sheets and Compu-Clearing ... 105

Figure 10.2: MPISI’s utilization level from load sheets and Compu-Clearing ... 105

Tables

Table 3.1: Safcor Panalpina and MPISI vehicle fleets ... 20

Table 7.1: Before and after the modification ... 56

Table 7.2: Safcor Panalpina’s transportation charges ... 58

Table 7.3: Transportation costs, expenses, revenue and profit for vehicle number F412 ... 59

Table 7.4: Illustrates the costs for each Safcor Panalpina vehicle ... 60

Table 8.1: Illustrates an example of the attributes that the cargo entity will receive from the excel file ... 62

Table8.2: Illustrates an example of the information that the vehicle entity receives from the excel file ... 63

Table 9.1: The difference between the actual payload and the simulation payload for SAPA ... 71

Table 9.2: The difference between the actual payload and the simulation payload for MPISI ... 71

Table 9.3: Illustrates the travelled distance in September, 2007 ... 72

Table 9.4: Illustrates the result of vehicle 889, day 2 ... 73

Table 9.5: Illustrates the actual cargo that has been loaded onto the vehicle 889 ... 73

Table 9.6: Difference between the original scenario and scenario 4 ... 86

Table 9.7: Comparison between the original scenario and scenario 6 ... 90

Table 9.8: Cost results for the original scenario ... 92

Table 9.9: The actual cost for the SAPA fleet ... 92

(13)

10

Table 9.11: The total cost for MPISI as well as Safcor Panalpina in scenario 1 ... 93

Table 9.12:The cost results from scenario 2 for the SAPA fleet ... 94

Table 9.13: The cost results from scenario 2 for the MPISI fleet ... 94

Table 9.14: The cost results from scenario 3.1 for the SAPA fleet ... 95

Table 9.15: The cost results from scenario 3.1 for the MPISI fleet ... 96

Table 9.16:Illustrates the cost results if F817 and F818 are replaced by MPISI vehicles in scenario 3.2 ... 96

Table 9.19: Illustrates the cost results if F412 and F817 are replaced by MPISI vehicles in scenario 3.4 ... 97

Table 9.21: The cost results from scenario 4 for the SAPA fleet ... 98

Table 9.22: The cost results from scenario 4 for the MPISI fleet ... 99

Table 9.29: The cost results from scenario 6 for both the SAPA and MPISI fleet ... 102

Table 10.1:Cost results and savings for each scenario ...107

Table 10.2:Minimum payload level ... 109

(14)

11

1 INTRODUCTION

The introduction clarifies the background and content of the thesis and illuminates the problems where to focus. Additionally, it describes and attains the objectives in order to improve the results and finally explain the essential limitations being made.

1.1 Background

From a logistics perspective, the South African transportation system is different from other countries in more than one way. Most of the cargo trades from foreign countries,

approximately 60 percent of the total consumption, is taken place the Gauteng region

according to Anthonie Verploegh1. This is rare because the majority of the import and export distribution is generally linked to a port. However, Gauteng is a province located in the inland of South Africa and includes large cities such as Johannesburg and Pretoria. Hence, there is no port linked to this province. The remaining percentage of the distribution mix is divided between the largest ports Western Cape, Port Elisabeth and Durban. Inside South Africa, road vehicles is the most common way to transfer cargo due to unreliable, insecure and deficient development of the railway. Other aspects of the South African transportation system to take into account are safety and long distances. In fact, many transport companies nowadays do not allow their drivers to transport the vehicles in the dark due to security issues. This leads to a reduced time window for travelling.

As a result of the highly exploited road network and increased demand from clients, the forwarding agent Safcor Panalpina’s intentions are improving their own growth in the same manner. Consequently, Safcor Panalpina (SAPA) expanded their warehouse area in August 2007 by opening a new facility called Phase 2 adjacent to the International Airport of Johannesburg. Phase 2 provides logistic solutions to clients and operates both as a terminal and storage function. To obtain efficiency throughout the whole supply chain requires more than a well performed facility. However, the distribution part is equally important. Safcor Panalpina considers the vehicle situation to be ineffective and would like an investigation concerning the subject.

1.2 Presentation of the problem

Safcor Panalpina’s aim is to constantly improve their business and increase their profit. A plausible solution would be examining the vehicle utilization in order to enhance the efficiency.

In the present situation, Safcor Panalpina is able to supply clients by their own vehicle fleet or by an outsourced company called MPISI. Due to an agreement, MPISI will be charged to deliver minimum 1 050 tonnage per month, even if the total volume delivered is less certain months. Safcor Panalpina assumes the utilization for all vehicles and both fleets are

unacceptably low. Since SAPA intend to reduce their own fleet, in order to expand the relationship with MPISI, they want to find the right mixture of vehicles economically viable for both parties.

1

(15)

12

1.3 Objectives

The purpose of this project is to study the current cargo outflow from Phase 2 and investigate the utilization level for Safcor Panalpina’s own vehicles and MPISI vehicle fleet. The

transport utilization for each vehicle will be analysed in order to find the optimal combination of vehicles and reduce the Safcor Panalpina fleet if possible. If the investigation indicates a low utilization level, the aim is to reduce the transportation costs. Hence, transportation costs for the present situation and for the new situation with accurate vehicle mixture will be

calculated. If the level is high, the goal is to find a feasible solution to smooth out the curve of “actual tonnage delivered by MPISI”. In other words, offer MPISI a more even volume each month if possible in order to strengthen the cooperation. The effort to achieve the objective is convened by two main questions and a few sub questions:

• What vehicle utilization level do Safcor Panalpina and MPISI currently have? o_{If utilization level is relatively high, find a way to smooth out the curve.} o_{If utilization level is low, find optimal mixture of vehicle fleet.}

§ What is the new optimal utilization level for all vehicles after the investigation?

• What are the transportation costs for Safcor Panalpina today? o_{What cost savings can be achieved for the new vehicle fleet?}

These questions will be answered during the course of the report and in the end, analysed and discussed.

1.4 Delimitations

Due to the complex problem of this study, limitations need to be done in order to reply to the purpose and achieve the aim within the given time frame. Other limitations and general assumptions will be discussed throughout the report and modelling assumptions will be described in chapter 7 “Data collection and assumptions”.

This study is limited to solely investigate the vehicle fleet of Safcor Panalpina and MPISI within the Gauteng region. The Gauteng region is a province in South Africa which comprises Johannesburg and Pretoria. Safcor Panalpina uses third party solutions for deliveries outside the region and therefore it appears natural to study only the SAPA and MPISI fleets in the Gauteng region.

Phase 2 is divided into two departments, logistics warehouse and degroup facility. Facility Phase 1 operates in the building next to Phase 2 and delivers cargo primarily to high value clients such as Hewlett Packard and Nokia Telecommunication. The Logistics warehouse stores cargo for various times and degroup facility handles cargo that needs to be delivered to client shortly after arrival to Phase 2, often within one day. This project is limited to examine degroup facility as these cargos are carried out on a daily basis and require fast deliveries. Since Safcor Panalpina’s own vehicle fleet is used at both Phase 1 and degroup, it is necessary to take Phase 1 into consideration during the study. However, focus will mainly be on Phase 2 in this thesis since most of the cargo is handled there. Furthermore, information and

documentation flow will not be a part of the study, only the flow of cargo from the dispatch part of degroup facility as well as from Phase 1 to the client will be considered.

(16)

13

1.5 Outline

Chapter 2 gives a short presentation of Safcor Panalpina and its connection to the

international supplier of forwarding and logistics services Panalpina and the carrier MPISI. If the reader is already acquainted with the information, he or she can easily skip this chapter entirely. The purpose of chapter 3 is to provide a surveyable representation of the present situation at Safcor Panalpina, e.g. how the flow of vehicles and cargo is operating. For the readers that are aware of the situation, it is possible to omit this chapter as well and continuing directly to the next section. Chapter 4 contains the theoretical main focus of the study and underpin the strategically and operational decisions that will be taken further on in the study. Information from here is based on actual facts from the literature regarding the subjects and is an essential factor of the thesis. This section, frame of reference, is well recommended to read in order to understand the remaining parts of the document. Moreover is chapter 5 dealing with the procedure of the project and includes each step in the thesis, from the beginning to the end. In addition are input data with regards to the literature included in this section, method. With the aim of understanding how the procedure from data collection to a finished project will proceed, it is reasonable to consider reading this chapter.

Next section, chapter 6 is an important part of the document since it describes the structure of the conceptual model. Furthermore, chapter 7 is mainly about the data gathered to the model as well as other data collected. The recommendation is to read these sections carefully because of the importance, especially if the reader is interested in understanding chapter 8, which explains the underlying structure of the simulation model.

Moreover is each simulation scenario and validation as well as verification of the model described in chapter 9. Further on, the chapter is presenting the results out from the scenarios. In other words, the utilization level for each vehicle and scenario and also the cost result. The main task for chapter 10 is to answer the questions asked in the introduction chapter. At least, chapter 11 presents recommendation for the future. The three last chapters 9, 10 and 11 are the most essential one for the reader that only is interested in the results of this thesis.

(17)

14

2 ABOUT SAFCOR PANALPINA

This chapter provides a presentation of the company Safcor Panalpina and its relationship to Panalpina and MPISI.

2.1 Safcor Panalpina

The information on Safcor Panalpina is based on an oral presentation by Greg James2 and the annual report (2006). Safcor Panalpina is a distribution company positioned in South Africa that provides supply chain management solutions to client worldwide and throughout South Africa. The company has 1200 employees in total and consists of 11 offices located all over the country. The main focus for Safcor Panalpina is movement of cargo and their brand promise stands for:

“A global supply chain of seamless motion”

In the year 1910 an enterprise called Fred Cohen Goldman & Cofound was established, which is the forerunner of Safcor Panalpina. Several decades later, in 1987, Safcor Panalpina

became listed on the Johannesburg Stock Exchange as Safcor Limited and was purchased by Bidvest Group Limited 1993. However, in year 2000 Safcor and Panalpina further integrated their business by creating a new company “Safcor Panalpina”. In essence, the main focus is products for the industry market, such as Automotive, Hi-tech, Pharmaceutical, Chemical, Mining, Telecommunications, Retail and Fashion. As mentioned earlier, Safcor Panalpina provides supply chain solutions that are divided into six different sections:

1. Supply chain management and consultancy 2. System integration

3. Customs clearing 4. Forwarding 5. Logistics

6. Financial services

Safcor Panalpina handles both air and sea freight to distribute to the clients. Some of their air freight clients are Hewlett-Packard (HP), Nokia Siemens Network, Siemens Limited, IBM, ABB, Schneider Electric and Toyota.

The organization is divided into four decentralized regions: Gauteng, Eastern Cape, KwaZulu Natal (KZN) and Western Cape. This thesis will be focused on the Gauteng region, a province in the north east part of South Africa. The largest cities are Johannesburg and Pretoria and the region consist of approximately 8 million inhabitants. All of the 500 employees of Safcor Panalpina in Gauteng are located in Johannesburg, divided into five different facilities: Phase 1, Phase 2, Skietlood warehouse, Wrench warehouse and at the ACSA building. The billing 2005 in the Gauteng region was R 5 418 billion per annum and Safcor Panalpina’s profit in the same region was 47 percent of the total profit according to figure 2.1. The Gauteng region handled more than 100 000 shipments and approximate 54 500 tons the same year.

2

(18)

15

Figure 2.1: Profit for each region

2.1.1 Relationship to Panalpina

Panalpina operates globally in the forwarding and logistics service sector based in more than 90 different countries all over the world. They are the second largest air freight provider and fourth largest worldwide to handle ocean freight and they employ 15 000 people all over the world. To provide national wide door-to-door solutions, Panalpina works closely integrated nation-wide with some partners and one of their partners is Safcor Panalpina.

Safcor Panalpina acts as an agent for Panalpina and they have established a close integrated relationship during the past 35 years. Hence, Safcor Panalpina uses Panalpina in their company name. Despite the dedicated relationship, Safcor Panalpina does not work as a subsidiary company to Panalpina. On the other hand they are a wholly owned subsidiary to Bidvest Group Limited and listed at the Johannesburg Stock Exchange. In contrast, Panalpina is listed at the SWX Swiss Exchange since 2005.

2.1.2 Relationship to MPISI

Black Economic Empowerment (BEE) is a common used term in South Africa introduced by the government with purpose to help BEE companies to grow. BEE companies are business that is black owned aimed at promoting black ownership of companies. The majority of a BEE company, more than 50 percent, must be black owned to qualify.

MPISI is an outsourced BEE company and operates on behalf of Safcor Panalpina. In other words, SAPA helps them continue to grow. MPISI acts as a transport company and provide Safcor Panalpina with distribution services, such as vehicles and drivers.

Due to a contract (Safcor Panalpina MPISI trading, 2005), MPISI is guaranteed to deliver minimum of 1 050 tons each month for SAPA. The contract reads:

“This agreement is concluded between Safcor Panalpina and MPISI on the understanding that the Drivers of MPISI will provide Distribution Services. Notwithstanding this MPISI shall remain ultimately liable and responsible for ensuring the fulfilment of all obligations and exercising and enforcing all rights arising pursuant hereto.”

(19)

16

3 PRESENT SITUATION

This chapter provides a detailed explanation of the present situation and will hopefully give a foundation for further understanding. Firstly, Safcor Panalpina’s facility Phase 2 is explained briefly. A more detailed description of the cargo flow from airport to client is included,

mainly concentrated on outflow from Phase 2 to client. The last chapter section clarifies the Safcor Panalpina and MPISI vehicles, both from a volumetric and kilogram perspective and number of vehicles. All information is based on oral references (c.f. list of references).

3.1 Phase 2

In august 2007, Safcor Panalpina completed and opened a new facility called Phase 2, nearby to the International airport of Johannesburg. The facility occupies a 10 000 m2 logistic area operating in two different ways. The first division of Phase 2, degroup facility operate as a terminal where cargo is received, broken down into pallet quantities, dispatched and delivered to the client. The second division of the facility, logistics warehouse is able to store cargo for a short or long time period and include the:

• Fridge area, for cargo in need of controlled temperature.

• Dangerous cargo area, to secure cargo that is dangerous to the environment, such as acids.

• Clean room, for cargo that requires a controlled level of contamination. • High value area, to store high value cargo, for example cell phones.

• Bound store, is the area for cargo that will be stored during a long period of time. However, the cargo can also be placed in the bound store if it is not customs cleared yet.

The reason why Phase 2 was built is because of the limited floor area in all other facilities. One of them, Phase 1, is situated next to Phase 2 and is shown in figure 3.1.

(20)

17

Figure 3.1: Phase 1 and Phase 2 location

Phase 1 was built two years before Phase 2 for especially two clients, Siemens and Hewlett-Packard. Nearly 60 percent of the 10 000 m2 logistics area is outsourced to Siemens and the remaining is handled by Safcor Panalpina and occupied mainly by Hewlett-Packard and some other high-tech clients.

3.2 Flow of cargo from airport to client

This section is mainly describing the cargo flow from Phase 2 to client. A short explanation of each step will follow in order to get a comprehensive view of the flow all the way from

airport to client.

Initially, all cargo is received at the Johannesburg International airport by air. Every week approximately six Boeing 747 flights arrive with cargo and each air freighter can carry up to 100 000 tons per flight. Safcor Panalpina allocates their cargo and use own vehicles to

transfer it from airport to Phase 2. The airfreight at this stage in the supply chain is packed in Unit Load Devices (ULDs). Cargo movement can either be done air side or land side

depending on the size of cargo. Air side means that collection by vehicle directly inside the airport is possible and cargo will be transported by a private road straightforwardly to Phase 2, which minimizes the risk of being hijacked or involved in an accident. Land side implies that cargo needs to be collected outside the airport area by a handling agent and transported on the public freeway to Phase 2. Once the vehicle has arrived to Phase 2, the movement of cargo starts inside the facility and will be moved forward in different steps. The receiving part starts during the unloading process of the cargo. In the receiving area, all cargo are checked and opened to inspect possible damages and further on split from ULD into pallet quantities. Certain clients demand their cargo to be weighted to ensure actual weight is correct and nothing was lost during the transport. The reason for this is because pieces of cargo are sometimes stolen, but at the same time the unit seems to be sealed. Consequently, some cargo is placed in queue to be weighted after the ULD split in the receiving area. In the next step, cargo is collected from the receiving area and transferred either to the logistics warehouse or

(21)

18

the degroup storage area where cargo are organized into lanes. Cargo is collected from the receiving area and allocated to the correct lane according to document number on the cargo. The next step, dispatch part will be described more detail in the following chapter subsection, as well as the last step transportation from Phase 2 to client.

3.2.1 Dispatch at Phase 2

The dispatch part is the last step of cargo flow at Phase 2. Cargo is collected from a lane and moved to a cage where vehicles are dispatched. A cage is an area inside Phase 2 from which the vehicles are loaded. There are nine cages and consequently nine loading places in the dispatch area. Depending on the geographical location of the client, the cages are divided into delivery areas to facilitate routes based on region. This prevents scattered routes all over the Gauteng region and avoids long distance between the clients’ locations. Each cage is

responsible for certain clients with different delivery areas belonging to them and each SAPA or MPISI vehicle is dedicated to a specific cage. The dispatch gets a trigger by printer, a draw sheet3, when cargo is approved to be loaded. The draw sheet indicates all documentation is completed and cargo is released by customs clearance. The cargo will be transferred from the correct lane to the correct cage depending on delivery area and further on loaded to a vehicle. When the vehicle arrives to the cage, dispatch staff coordinates cargo in cage and decide which clients to visit during the route due to client delivery area and quantity of cargo. Figure 3.2 shows an overview of the cargo flow from Phase 2 to client, where cargo arrives to the receiving area, is transferred to the correct lane, allocated to the right cage, loaded on a vehicle and finally transported to the client.

3

(22)

19

Figure 3.2: An overview of Phase 2 and the cargo flow from Phase 2 to client

3.2.2 Transportation

When the cargo is loaded on the vehicle the transportation process begins. In other words, cargo moves from Phase 2 to client. According to Jacques Erasmus4, the dispatch staff will

4

(23)

20

always try to load the vehicle full, but the cargo handled by Safcor Panalpina fluctuates in sense of volume, actual weight and shape. Although the volumetric size of the cargo is large, the weight can be low, which makes it difficult to load the vehicle with high utilization level from a weight perspective. Volume and weight will be described further on in the report, in chapter 6 and 7. A vehicle can be loaded one or several times per day, depending on distance to clients, road traffic congestions, unloading time and loaded quantity to each client. In some cases, a vehicle is fully loaded to only one client. If one vehicle is fully loaded to one or several clients and unloading time and distance to and between clients are short, the vehicle will probably be able to load a second time or maybe even a third time in one day.

Transferring cargo from Phase 2 to client can occur in four different ways: 1. By Safcor Panalpina vehicle fleet.

2. By MPISI vehicle fleet

3. Client collect cargo from Phase 2

4. By outsourced transport companies demanded by client. For example Berco, Schenker and DHL.

3.3 Safcor Panalpina and MPISI vehicle fleets

There are four ways to deliver cargo to client, but this thesis will only focus on two of them, Safcor Panalpina and MPISI vehicle fleet. According to Winston Mullany5, most of the general cargo is delivered by MPISI fleet in the present situation, but some clients demand to be delivered by SAPA fleet. Hence, Safcor Panalpina’s own vehicles deliver mainly to clients with certain requirements.

All vehicles are shown in table 3.1, including the number of vehicles in each fleet, loading weight and volume capacity. Safcor Panalpina owns totally 11 vehicles, mainly horse and trailers. MPISI owns almost entirely 8 tons vehicles and provide Safcor Panalpina with 14 vehicles in total. Small vehicles are called “bakkies” and are able to load maximum one ton.

Table 3.1: Safcor Panalpina and MPISI vehicle fleets

Company Vehicle Number Payload capacity (kgs) Volume capacity (m3)

SAPA 1 ton 2 1 000 4 SAPA 4 tons 3 4 000 16 SAPA 8 tons 5 8 000 47 SAPA 18 tons 7 18 000 81 17 180 000 MPISI 4 tons 2 4 000 16

MPISI 8 tons (closed body) 8 8 000 42 MPISI 8 tons (curtain side) 2 8 000 42

MPISI 8 tons (Opened) 2 8 000 Greater than above

14 104 000

Total 31 284 000

5

(24)

21

4 FRAME OF REFERENCE

The frame of reference will initially describe a definition and provide a brief summary of the terms logistics and transportation in order to illuminate the reader background theory related to the subjects. In addition, it will clarify the connection between delivery service and

profitability as well as the relationship between logistics and transportation. Hence, describe how to calculate the utilization level for the vehicles. Furthermore, it will explain the main concept “operations research” and clarify the different operations research methods

“simulation”, “optimization and “heuristics”. This provides the reader with knowledge and understanding of why one of these methods will be chosen in the next chapter.

4.1 Logistics

According to Council of Supply Chain Management Professionals (CSCMP, 2007) the definition of Logistics Management is:

Logistics management is that part of supply chain management that plans, implements, and controls the efficient, effective forward and reverse flow and storage of cargo, services and related information between the point of origin and the point of consumption in order to meet clients' requirements.

In other words, the objective for logistics is to achieve cost efficient delivery service.

According to Aronsson et al (2004) logistics is more about doing the right things than doing things right. Logistics involve two main goals, achieve as high delivery service as necessary in order to satisfy clients and at the same time provide client with as small logistic total cost as possible. To achieve the goal for logistics all the employees need to endeavour to observe the complete flow instead of all employees or departments attempt to fulfil their own goals. The old approach is called the functional way of thinking where the company uses the individual and solitary way of thinking and tries to optimize each department flow. The logistics way of thinking is called process thinking where focus is on the entire flow and the objective function is to minimize the flow or the processes in order to get a comprehensive view of the entire organization. The difference between a process oriented company and a functional oriented company is that the former implies that there are distinct clients and the latter implies that the clients are indistinct. There are a few drawbacks with the functional way of thinking. All the problems that occur in the different departments have a tendency of summarize in the end. No one is responsible for the comprehensive view, the different departments hand over the responsibility to the next one. Each department’s economy follow-up system will only focus on their own department and not the entire company. This implies competition between the departments which will affect the profits negatively.

According to Waters (2006) the logistic costs consists of several activities and procedures such as transporting, receiving, warehousing, inventory management, material handlings, distribution, locations decisions etc.

4.1.2 Delivery service

According to Aronsson et al (2004) the definition of delivery service is to satisfy the clients demand on delivery service. In order to please the clients there are three delivery service activities that are important to understand:

(25)

22 •

••

• Before the delivery, the essential part is to let the client know what they can expect from the service and be able to handle clients requirements and make necessary adjustment to satisfy every clients specific needs.

• ••

• During the delivery, it is important to deliver in time, the right number of products and to the correct delivery address. If an ordered product fails to be delivered on the right day, it is important in an early phase as well as to continually inform the client about the current situation.

• ••

• After the delivery, involve mostly handling complaints, returning cargo and offering repair parts and at the same time act service-minded.

4.1.3 Relationship between delivery service and profitability

Achieving a good and acceptable service level requires a balance between delivery service and costs. The aim is to obtain the highest level of service possible to the lowest feasible cost. Delivery service can be measured by seven different parameters and the relevance of each delivery service element differs depending on the client. The seven different parameters are:

• •• • Lead time • •• • Delivery reliability • •• • Delivery dependability • •• • Information • •• • Customization • •• • Flexibility • •• • Warehousing availability

Figure 4.1 illustrates where the limit of service is no longer gainful to the company. The service level should be somewhere between 95 and 99 percent, everything above this level result in a loss of profit. If the service level is 100 percent it means the company will deliver to the client exactly at every occasion and end in an almost infinite cost. The marginal cost is the extra cost for increasing the service one additional level and the marginal cost increases heavily when the service level is too high. The company should, in other words, strive at finding and keeping a service level that achieves the maximum profit. Hence, if the service level is above the best suitable maximum level the cost increases at the same time as profit decreases and the revenue curve is levelled out. If the client needs to pay for the extra service by themselves, their will to pay usually decreases as the service and costs increases, which explains why the revenue curve is levelled out (Aronsson et al, 2004).

(26)

23

Figure 4.1: Relationship between the delivery service level and profit

4.2 Transportation

The essential part of the transportation is to move people and cargo from one location to another. This process generates different kinds of utilities to the client such as time and place utility. Time utility means that the transport service will be used only when it is necessary, for instance a product is stored in the warehouse during a certain time frame and will be

transported to client when there is a demand for the product. Place utility signifies that cargo will be moved from one place to another if the latter place brings larger utility to the process (Lumsden, 2006).

4.2.1 Relationship between logistics and transportation

Transportation does have a close connection to the logistics concept and involves movement of cargo. Each time a transport business needs to transfer material from destination A to B, a transport is needed which for instance means transfer between a factory and storage or between storage and a client. Transportation occurs in every step throughout all logistical processes in a company. However, transportations can be divided in internal and external transports. Internal transports transfer material within a limited area, for instance within a manufacturing or warehouse area. External transports mean moving material between

different areas, for example from one region to another region or between different companies in the same region.

There are three different perspectives and levels of transportations. The higher level is from the purchaser’s and owner’s of transportations perspective and implies the movement of cargo from one node to another. The purchaser demands the lowest feasible price, shortest lead time and highest delivery service for the cargo. In the earlier days, large quantities where ordered occasionally. Nowadays, it is actually more common to order small quantities frequently which results in more carrying. More regular transportations mean decreased storage levels at the client and increased demand for the transport company. Transport companies need to develop the transportation flow in order to manage and handle the new requirements.

(27)

24

Furthermore, some clients do even require on-time delivery and use time windows to prevent late or early deliveries. A time window implies the driver needs to arrive to the client within a specified time frame to be able to unload. This method is commonly used and applied by business companies working after a principle called JIT deliveries (Just-In-Time). The differences between traditional deliveries and JIT deliveries are many. The characteristics of the traditional deliveries are large quantities, no on-time deliveries, and long transportation time. JIT-deliveries on the other hand signify small quantities, on-time deliveries and short transportation time (Aronson et al., 2004). The latter alternative result in poor utilization level of vehicles and it is usually difficult to receive a high level of fill level in the vehicles due to the frequent deliveries (Lumsden, 2006).

The middle level describes the perspective of the transport companies’ owners and is often more complex than the higher level. This level is where cargo is usually consolidated with other cargo and they are able to be reloaded more than once during the transport. The transport market is constantly changing, all in order to satisfy the clients. Consequently, the transport companies need to follow the changes in the same manner.

The lowest level describes the traffic market that the owner of the infrastructure depends upon in order to perform deliveries. The infrastructure implies roads, railways and airports. This level is a requirement to make the middle level exist (Aronsson et al., 2004).

Waters (2006) lists a number of improvements that are necessary for the company in order to make the logistics operations more efficient. Some of the methods to apply are helpful tools to improve the business and are described below:

• ••

• Cross docking. This method coordinates the supply and delivery where the goal is to reduce and later on remove the entire warehousing area. When the cargo arrives to the cross docking, it is directly transferred to the loading area and thereafter loaded on to a vehicle. The cargo will not be stored, but there might be some sorting of cargo such as break bulking or merging.

• ••

• Direct deliveries. One way of reducing the lead time and the costs is to use direct deliveries, for instance buying through Internet or directly from the producer, since more and more of the clients want their products faster. One of the major advantages is the reduced supply chain where the clients are able to have direct contact with the manufacturers.

• ••

• Small deliveries. Direct deliveries and JIT consequently reduce the cost and lead time which leads to smaller deliveries. Transport companies such as DHL consolidates small deliveries which increase the utilization level for the vehicles and provide a more efficient supply chain.

• ••

• Increasing vehicle utilization. One way of making deliveries more efficient is to increase the vehicle utilization. This can be done in several ways such as using

backhauls (the vehicle finds loads on the way back from the route, which improves the fill level in the vehicle significantly), reverse cargo (cargo that needs to be repaired or recycled), freight forwarding (cargo from different clients are combined for the delivery) and more efficient schedules (change of routes for the vehicles).

(28)

25

4.3 Resource utilization of vehicles

According to Lumsden (2006), the resource utilization describes how businesses are able to use their resources most efficiently. For manufacturing companies this may imply using their machines as efficiently as possible and for the transportation businesses this might involve utilizing the vehicle fleet in the most efficient way from a capacity-, time- and velocity perspective. Since there is no significant difference in velocity between the vehicles due to a fixed maximum speed limit on the roads, consequently only two dimensions are left, capacity and time, which are presented in greater detail.

Considering the resource utilization of the vehicle fleet there are two levels of measuring the utilization, physical and economical. Physical utilization involves capacity and time

dimensions and signify the travelling utilization from location A to B, but do not include the empty vehicle return load to location A. With the intention of receive as accurate result as possible, the economical utilization should be taken into account while it includes the return load to location A. The resource utilization (Reu) for the loading capacity on each vehicle can be calculated in the following way (Lumsden, 2006):

Reu (loading capacity for weight or volume) =

capacity

loading

Available

capacity

loading

Utilized

The time utilization can be calculated in the same manner as Reu (loading capacity): Reu (time) =

time

Available

time

Utilized

In order to get the physical utilization level it is essential to consider both capacity and time: Reu (physical) = Reu (capacity) * Reu (time)

An example demonstrates the above mentioned formulas. If the maximum loading capacity is 18 tons for one vehicle and the actual loaded weight is 10 tons during one route, the capacity utilization is 55 percent. If the vehicle is used eight hours uninterruptedly, from 8a.m. to 4p.m, the time utilization is 33 percent of one day. Hence, the physical utilization level is

accordingly 18 percent, which is a relatively low percentage unit and that is why the economical aspect should be taken into consideration. The economical utilization level can therefore be calculated as below:

Reu (economical) = Reu (capacity) * Reu (time) *

time of unit Cost time of unit venue / / Re

4.3 Operations research

Operations research (OR) is a commonly used term to solve complex problems and consists of several different scientific methods and techniques in order to provide the decision maker with quantitative management information

(

NE, n.d.). Operations research is a valuable tool for planning, evaluating and analysing on an operational as well as tactical and strategic level. Quantitative or mathematical analysis is basically the initial point to make decisions for all operations research fields (Lundgren et al., 2004). Operations research and management science is often used synonymously, but management science is frequently associated with an

(29)

26

applicable method for business management problems. OR on the other hand operates in the industrial engineering field (NE, n.d.), (Thomas, n.d.). The concentration for this thesis will however focus mainly on the term operations research.

There is a widerangeof applications for OR. For instance, to find the most suitable alternative that generates the best result or maximizing the utility for all resources (Lesso, n.d.). In addition, there are algorithms for routing problems and methods for identifying and analysing capacity, utilization and bottlenecking problems (Thomas, n.d.).

The concept OR was developed during World War Two, when several groups of scientists were put together in order to analyse different defence methods (Lesso, n.d.) and military operations. One of the best known OR methodologies developed during this period was linear programming by George Danzig. When the war was finally over, operations research

continued to be applied in the military area, but the real breakthrough came in the early fifties when operations research started being exploited in the industrial area (NE, n.d.). For the first time it was possible to optimize problems instead of using the trial and error method. (Lesso, n.d.)

4.3.1 Three different approaches

According to Pidd (2003) the method can be divided into “soft” and “hard” approaches depending on the aim of output. Hard methods can be distinguished from soft methods by the decision making, control and business process approach. Soft models can be explained by assisting in strategic decisions and planning and is able to make a significant affect on the business. In essence, soft models make it easier for people to understand each other. Occasionally, people share the same experience and yet may interpret the reality very differently. Soft models deal with the fact to find a suitable language to enhance the knowledge for people and their experience of different situations.

The hard methods, mathematical and logical modelling, can be divided into three main areas and will be explained more detailed in the following sections:

• Computer simulation • Optimization modelling • Heuristic search

4.4 Computer simulation

A definition of computer simulation (Kelton et al., 2003) is:

Simulation is the process of designing and creating a computerized model of a real or proposed system for the purpose of conducting numerical experiments to give us a better understanding of the behaviour of that system for a given set of conditions.

The basic idea of computer simulation is to use a computerized mathematical model to imitate the behaviour of a real system with the aim of studying and analysing the output result

(Chung, 2004). A model is useful for experimentations and will answer the question “what if?” (Pidd, 2003).

(30)

27

Simulation is a useful tool in decision making situations and is able to provide information to help the decision maker to choose the most suitable alternative from a range of options with different criteria’s (Pomykalski, n.d.).

According to Chung (2004), simulation can be used for many different systems such as transportation, service and manufacturing systems. The fields of transportation systems imply for instance:

• Airport processes with departing passengers flow from queuing at check-in counters, further on to security checkpoint and all the way to the boarding procedure.

• Port shipping processes where the number of cranes and vehicles, to unload a vessel, can be determined.

• Train, bus and vehicle transportation to simulate routes.

• Distribution network to analyse facility location and transport links.

4.4.1 Simulation versus simulator

Simulation can be divided in two types of simulation models, computer simulation and computer simulator, where focus for this thesis is simulation. Both of them are used for physical or proposed systems and the models contain different resources or operating policies. However, resources and operation policy for simulation is determined beforehand and the decisions are made during the run for simulators. Simulation aids the user to compare performance from the output measures between different scenarios, to make the correct decisions concerning the system. Simulator, in contrast, is able to provide the user visual effects during the simulation run in real time and the aim is to train the user on how to make decisions rather than make own decisions (Chung, 2004).

4.4.2 Considerations before using simulation

There are several purposes of using simulation as a tool. The method is able to deal with rather complex problems and create substitute models of the real system with the purpose of studying without doing any harm to the system. However, simulation might help answering the question “what will occur in the system if I change one or more parameters?”, for example change the number of resources or if a manufacturing industry wants to implement a new machine to the facility. The user will also be able to try new ideas without affecting the real system (Kelton et al., 2003).

According to Chung (2004), processes and operations belonging to complex systems can be hard to understand without a dynamic model. To study the system by stopping it or

investigating certain components in the system separately can be difficult or even impossible. A representative example is finding were the bottlenecking problem takes place in a

production plant. Furthermore, simulation is used to improve already existing systems in order to develop different processes in the system. For instance, changes in break scheduling for all resources.

For some systems that do not yet exist there is an advantage to perform a simulation model before building a real system. The simulation model is able to provide directives in a good or bad way, in order to understand what the new system will accomplish. The cost for modelling and do experimental investigations before installing the system physical is low in comparison to construct the system in advance and do all testing afterwards. Additionally, when

purchasing and implementing new equipment to a manufactory it can be gainful to study if the system will operate as intended beforehand.

(31)

28

Considering systems where the level of sensitivity is high or so very critical that they cannot be disturbed, simulation models have the possibility to analyse these systems. Executing physical experiments for a critical system can cause serious impact on the system. Nuclear power plant is an example of a highly critical and sensitive system.

One major advantage by simulating the model on a computer is all the time saved compared from testing the experiments in a real system, which sometimes take months to accomplish. The time for all experiments will be compressed by the computer during the simulation runs. Furthermore, an increased statistical reliability for the analysis is possible to achieve by running the model for multiple replications. The development of computer simulation has made it easier to study and analyse complex systems dynamically during the simulation runs. In fact, this has resulted in a comprehensive range of different systems to analyse compared to the past where mathematics and programming skills were necessary. In addition, computer simulations possess the ability to present the model visual, which is unusual. Most operations research methods are only able to reveal numerical calculation and textual based

demonstrations. The simulation model has the option to be animated dynamically to enhance the analyse results. An animation is able to locate and remove defects in the system as well as handle different situations and is a useful tool to demonstrate the behaviour of the system and how it works.

Despite of all benefits, there are a few disadvantages by using simulation. First of all, it is impossible to achieve accurate output data when the model does not have accurate input data. Data collection is the most critical item of the process due to the complex reality and often unreliable input data. Furthermore, complex problems lead to complex answers. If a complex problem have a large number of parameters to take into consideration, like most of them do, it can be necessary to simplify the problem and do certain assumptions. Otherwise, the

simulation model can be unreasonably complex and extensive to solve within a limited time period.

Other considerations to take into account before building a simulation model is the time perspective, uncertainty perspective and knowledge on the subject of statistics and modelling. Developing and building a model will presumably be time consuming, especially when developing a simulation model from a complex system. Data collection, model building and analysis are three steps in the modelling process that requires plenty of time (Chung, 2004). Another approach to take into account is the uncertainty in the output result, in particular when the modelling time frame is relatively short (Kelton et al., 2003). In order to simplify complex modelling, reduce time and decrease uncertainty, modelling assumptions will be needed, Hence, it is important to make adequate and fair assumptions, otherwise it will probably injure the model and result in invalidity (Chung, 2004). It is impossible to measure or reduce the errors of uncertainty from assumptions in the invalid model. However, it is better to receive an approximate answer to a correct problem than an exact answer to the wrong problem (Kelton et al., 2003).

At the present time, it is far simpler to create a simulation model and analyse the output result than it was in the past, but at the same time notably complicated anyway. In the earlier days, there were no graphical interface, only text based systems and practitioners had to do all the computer programming by themselves. Nowadays, it is still a difficult process and it requires knowledge to understand the system correctly in order to interpret the simulation model in the

(32)

29

right way (Chung, 2004). Since many real systems are stochastic, random input data is required in most simulation models and there cause random output (Kelton et al., 2003). Therefore, most of the simulation models include statistical techniques and hence statistical knowledge is needed to analyse the result (Chung, 2004).

4.4.3 Classification of simulation models

A simulation model can be classified into three different dimensions; the state of the system, randomness and time. The categorization affects which kind of simulation tools will be used. When changes occur continuously over the time, the model has a continuous state. Systems that actually appear at specific points in time and where nothing takes place between two events are called discrete models. Additionally, randomness is an important factor when modelling. When there is no random input variable involved (Kelton et al., 2003) and when events occur in a predictable pattern or in the same way each time, the model is called

deterministic. Deterministic data needs to be collected only once since there will be no change of values throughout time (Chung, 2004). For instance, a manufacturing machine in process is rate based and will always take the same time to run. In contrast, stochastic models always have at least one random input parameter to take into consideration. For example, the route for a vehicle loaded with cargo do not always take the same time to perform, depending on the traffic congestion and capability of driver. In other words, deterministic models are affected by computers and will always take the same time to execute. Stochastic models are affected by the human hand and will therefore cause varying time to run. Furthermore, the most essential factor for dynamic models is time. In fact, nearly all operational models are dynamic. If time does not play an important part on the other hand, the model is static (Kelton et al., 2003).

4.4.4 Procedure in a simulation study

Considering the procedure during the simulation study, there are a few step which are essential to follow in order to improve the quality and solution of the study according to Kelton et al (2003) and Chung (2004):

• To get an overview of the real system and understand the problems. The first step can be reached by interviewing the people working in the real system. That gives a deeper understanding of the situation.

• Specify the objectives already from the beginning. An essential part of the study is to make the objectives clear and not promise anything impossible to solve.

• Gathering input data and analyse it. One of the most important and difficult step is the data collection since the output quality reflects the input data. Without proper data there will be an unreliable result.

• Create a conceptual representation of the system and formulate it into a model. This can be done by building a flowchart of all the processes linked to each other. It is furthermore important to specify the appropriate level of details and decide which assumption needs to be done. A low level of detail will probably generate poor quality of the output and too high level of detail can lead to a complicated model.

(33)

30

• Translate the conceptual model into software. The model needs to be translated into simulation software. There is a wide range of simulation software available e.g. Arena and ED, where the objective is to find the most suitable one.

• Verify that the model is built correctly. Verifying that the conceptual representation corresponds to the computer model in a proper way and make sure it is an accurate model. This can for instance be met by using animation. Animation provides a visual perspective of the model and facilitates the trouble shooting.

• Validate that the correct model is built. This involves the procedure of comparing the computer model versus the real system and control if it performs properly. Both verification and validation are vital steps in a simulation study.

• Planning and run the experiments. Design and planning is the first step in the

experimental phase and involves planning of which experiments to be executed. When the design is made and the experiments are defined, then it is time to run the

experiments. Probably, this procedure will take some time for the computer to solve. It is important to determine the conditions before the run, for example how many times to run the simulation to achieve accurate output.

• Analyse the output results from the experiments. Examination and analysis of the results can be performed by using a statistical tool, if the system is stochastic. • Evaluation of the results and to get insight. This last step is necessary in order to

understand if the results make sense. Without plausible results it will be hard to guarantee the reliability of the model.

4.4.5 Verification and Validation

Verification and validation are two iterative processes that should be carried out continuously and separately throughout the simulation modelling phase. They are both important steps in the simulation procedure in order to present and enhance the reliability of the result (Chung 2004).

Verification is the process to examine if the model operates as intended. In other words, comparing the simulation model to the logical processes and activities in the conceptual model with regards to the modelling assumptions made in an earlier stage. If the system is extremely large and comprehensive, it may be hard to ensure that all the logical steps and data in the model are correct. It is usually easier to handle smaller systems. To achieve a successful verification there are a few techniques that should be taken into consideration. One common technique is to follow a single entity throughout the system in order to ensure the system behaves as planned.

By using constant variables in the model, i.e. deterministic data, instead of the original random inputs helps to foresee if everything in the model works as intended (Kelton et al., 2003). Another way of verifying the model is to break down a large model to several sub models and debug those one by one, a method called “divide-and-conquer”. This facilitates the process of detecting faults in the system. Once the smaller models run correctly, it is possible to improve the output for them. For instance, including new essential components at different models in order to reflect the actual system in a truthful way. Other techniques of ensuring the model works properly is by using animation or manually advancing the