Sustainable Cities and Society 69 (2021) 102840
Available online 9 March 2021
2210-6707/© 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
A mixed method evaluation of economic and environmental considerations
in construction transport planning: The case of Ostl¨anken
Anna Fredriksson
a,*
, Pamela C. Nolz
b,c, Cl´ovis Seragiotto
baLink¨oping University (LiU), Department of Science and Technology, Communication and Transport Systems, Construction Logistics Group, 601 74 Norrk¨oping, Sweden bAIT Austrian Institute of Technology, Center for Mobility Systems, Dynamic Transportation Systems, Giefinggasse 2, 1210 Vienna, Austria
cSt. P¨olten University Of Applied Sciences, Matthias Corvinus-Straße 15, A-3100 St. P¨olten, Austria
A R T I C L E I N F O Keywords:
Infrastructure construction Environmental impact Construction transport efficiency Soil and rock materials Optimization Case study
A B S T R A C T
Among the construction related transports, the transport of soil and rock materials stand for a major part. The purpose of this study is to develop an approach enabling scenario analysis of the relation between storage location and soil and rock material transport planning. The study follows a sequential exploratory mixed methods design. First a literature review and a qualitative case study identifies the problem. Second a quanti-tative optimization method is used to evaluate possible scenarios showing the interdependence between storage location costs and transport impact. The study has two main contributions, the mixed methods approach to evaluate economic and environmental considerations in an infrastructure project and the scenario analysis of different options for inventory control and transport. The presented study adds knowledge to transport efficiency of rock and soil materials.
1. Introduction
The construction industry is not only suffering from low productivity growth but is also responsible for about 40 % of the energy use and one third of greenhouse gas emissions (GHG) generated world-wide (Pearce & Ahn, 2017). Decreasing the environmental impact of the construction
industry is a current and an immensely important issue. Even though the concept of sustainable construction is not new, improving the applica-tion is still in its early phases (Ozcan-Deniz & Zhu, 2017). Earlier studies have mainly focused on promoting carbon reduction technologies (Du et al., 2019) or measuring GHG emissions (Sandanayake, Luo, & Zhang, 2019). Though, the potential of new planning processes should not be forgotten (Sobotka & Czarnigowska, 2005) and how this can be trans-lated into new processes and policies (Du et al., 2019).
Xu, Deng, Shi, and Huang (2020) show that CO2 emissions from
construction material transports are a substantial problem. According to
Seo, Kim, Hong, and Kim (2016) construction transport accounts for 2.4 %–5.5 % CO2 emissions, whereas according to Chang and Kendall (2011) it is up to 16 % of the emissions in a project. Among these transports, the transports of excavation materials, i.e. soil and rock, need special attention. Goods transport surveys show that these form the largest part of the ton-km transported in a Swedish region (Dalenstam,
2015). They travel short distances; however, because of their number their impact is large (Eras, Guiterrez, Capote, Hens, & Vandecasteele, 2013).
Transport efficiency is essential to reach Sweden’s goal of fossil free transport in 2045 and UN SDG13 (taking urgent action to combat climate change and its impacts). McKinnon (2018) presents seven fac-tors that limit the transport efficiency, all related to planning: Trade-offs between cost and delivery service/inventory costs; Lack of information regarding available transport capacity and capacity demand; Lack of redundancy in transport scheduling decreases the possibility of bundling; Low fill rates because of empty space due to incompatibility between packaging and transport dimensions; Functional silos and lack of cooperation; Imbalances in traffic flow; and Regulations. Therefore, the potential to increase transport efficiency through improved planning is great (Eriksson, 2019), which is also true for soil and rock materials.
Magnusson, Johansson, Frosth, and Lundberg (2019) conclude that strategically located storage locations, i.e. inventories, for material co-ordination could reduce the demand for soil and rock transportation by 23–36 % per site compared to business as usual. Eras et al. (2013) state that improved handling of soil and rock materials can reduce GHG emissions from these transports by up to 41 %. Thus, one way to decrease the environmental impact of infrastructure projects is by
* Corresponding author.
E-mail address: anna.fredriksson@liu.se (A. Fredriksson).
Contents lists available at ScienceDirect
Sustainable Cities and Society
journal homepage: www.elsevier.com/locate/scs
https://doi.org/10.1016/j.scs.2021.102840
vated; type of ground engagement gear fitted that affect the fuel and thereby the GHG emission in a project. However, these have been left outside this study to enable focus on how the environmental impact can be decreased through improved coordination of the transport and stor-age area planning. When the results of the study are to be implemented these factors of course have to be considered.
The contributions of this paper are twofold:
1) With a mixed methods approach a way to decrease the environ-mental impact of infrastructure projects is presented, by improved planning of the transport and storage areas for rock and soil mate-rials. Problem features are investigated qualitatively, while the impact of action scenarios is evaluated with a quantitative optimi-zation approach.
2) A scenario analysis provides managerial implications of how to both decrease environmental impact of construction transport and improve construction productivity, i.e. considering a longterm planning horizon and aiming for high fill rates in vehicles that optimally use transport resources. The scenario analysis enables the developers and local authorities to understand their action space by visualizing the trade-offs, which are encountered in the planning processes.
The rest of the paper is organized as follows. First, the literature review is presented, including a description of rock and soil materials specifics, construction transport planning and inventory control. Sec-ond, the problem is described in detail. Next, the research design is introduced explaining the mixed methods approach. Thereafter, the case study is presented followed by the experimental results. Finally, con-clusions and recommendations are brought up.
2. Literature review
This section first describes the specifics of rock and soil materials transport and handling, thereafter the current state of research within construction transport planning and finally earlier studies of modelling inventory control problems.
2.1. Excavated materials transport and handling
Excavated rock and soil are bulk materials (Woodcock, 2015) and can also be classified as a type of solid waste (Derrible, 2018). Typical sources are construction sites, road repair, renovation and demolition of buildings (Derrible, 2018). The handling of solid waste includes six functional elements (Derrible, 2018): 1) generation, 2) handling, sepa-ration, storage and processing at source, 3) collection, 4) transfer and transport, 5) separation, processing and transformation at another location, and 6) disposal. The transfer and transport step involve many operations, though the main are the loading of the truck in a fleet of trucks, the haulage of the loaded material to a storage area, the return of the truck to the loading area (empty or loaded with material to fill)
2.2. Construction transport planning
There are specific actions that can be taken to improve the efficiency of the transport system (Aronsson & Huge-Brodin, 2006; Martinsen & Huge-Brodin, 2014; McKinnon, 2018):
- How the system is organized on a structural level - Transport planning
- Fill rates
- Return transports – empty transports - Change of transport mode
Akbarnezhad and Xiao (2017) identify from a literature review that the main factors impacting the rock and soil material transport emis-sions are the quantity of material to be transported, the size of the ma-terial, the transport distance and the mode of transport. Bulk materials are in most cases heavy, and are therefore in most cases dealt with on a local market and the bulk trading companies are organized based on a geographic basis in order to minimize transport distance (Woodcock, 2015). The fuel cost of soil and rock equipment is also a substantial part of the project cost and to decrease the used fuel would be beneficial not only for emissions but also for project costs (Kaboli & Carmichael, 2016).
Rock and soil transports are not independent from the rest of the logistics system and actions taken in other parts of the logistics system can have a large impact on the number of transports as well as their transport efficiency. These transports must share the infrastructure with other road users (Behrends, Lindholm, & Woxenius, 2008; Dablanc, 2007) and the additional transport flows induced by construction create a conflict situation regarding the city infrastructure. In a city, different roads have different capacity as well as there are different levels of sensitivity. Sensitivity means if e.g. schools or hospitals are passed on the way to and from a construction site. A natural action is to try decreasing the number of construction transports or at least affect their routing. Some urban areas have restriction policies on weight and di-mensions of vehicles. Sometimes these policies can have the opposite impact compared to their intention, i.e. they increase number of trans-ports as fill rates have to be decreased to keep truck and load weight within limits (Treiber & Bark, 2016).
Transport planning in construction is dependent on the planning of construction works. In a construction project the main contractor is responsible for the planning of the production (B¨ackstrand & Fre-driksson, 2020). (Aronsson & Huge-Brodin, 2006) show how the struc-ture of storage locations affects the environmental impact of transports. The location of rock and soil storage areas is part of site layout planning. Site layout planning is the process of identifying the number and size of temporary facilities, locations, sizes and routes (Sanad, Ammar, & Ibrahim, 2008; Song, Xu, Shen, & Pena-Mora, 2018). The site layout and the transports should be planned together, as there is a high level of interdependence between these (Song et al., 2018). However, this is not to its fully extent done today. The base for rock and soil materials
transports is the mass haul diagram (Smith, 1995). This diagram pre-sents cut and fill volumes of different areas of the project. The coordi-nation of excavated soil and rock within as well as between projects requires early planning (Magnusson et al., 2015). The knowledge about future resource needs and quantities and qualities of excavated soil and rock generated in a project is a basis for such planning (Magnusson et al., 2015). One possibility is to open up for regional management of the coordination of these materials (Magnusson et al., 2015). Presently, the planning of bulk transport in construction is made daily and lacks long term focus (Woodcock, 2015). Being characterized by heavy weight, fill rate and payload and these transports are in most cases maximized un-less a smaller load has been ordered (Woodcock, 2015). Utilization of vehicles is key, however there are efficiency losses due to time spent waiting to load or discharging and long distances travelled (Woodcock, 2015). Another common challenge are small storage locations and the general aim in the industry to start with a full stock in the morning and end the day with an empty stock, which creates peaks in transport de-mand and decreases utilization of vehicles (Woodcock, 2015). The loading and unloading of vehicles are highly dependent on the avail-ability of excavators. There are great potentials to improve efficiency and cost reduction by rationalizing these internal logistics processes (Woodcock, 2015). An important issue is the truck size. Kaboli and Carmichael (2016) see indications that a smaller truck size actually can decrease fuel consumption in a project as well as maintenance cost of roads.
According to Magnusson et al. (2015) the focus of earlier studies of excavated soil and rock has been waste management and recycling. Furthermore, earlier research within site layout planning has focused on either the site layout planning per se or the material transport planning per se (Song et al., 2018), not the relation between the two. Therefore, there exist several models of how to create an optimal site layout and how to enable the recycling of soil and rock masses within the project and between the projects. However, studies on the relation between transport and site layout planning (facility location and inventory allo-cation) for these materials are lacking.
2.3. Coordination of transportation and inventory
The relation between facility location, inventory allocation and transports has been studied and modelled before within operations research, though not based on the case of infrastructure construction.
Federgruen and Zipkin (1984) show the positive economic effects of the coordination of transportation and inventory decisions. As opposed to filling individual orders of material as they are placed, the quantity and timing of deliveries are coordinated and determined by the supplier. This supply-chain policy is called vendor-managed inventory (VMI), where the supplier manages product inventory at customers (i.e. con-struction sites) to ensure that customers do not experience a stock-out.
Cetinkaya and Lee (2000) underline that VMI systems minimize distortion of demand information (known as bullwhip effect) transferred from the downstream supply-chain member (e.g., retailer, construction site) to the upstream member (e.g., supplier). Considering shipment scheduling at the same time as stock replenishment, costs for the dis-tance travelled can be reduced, e.g. by consolidating several smaller shipments into a single load.
Burns, Hall, Blumenfeld, and Daganzo (1985) develop an analytic method for transportation and inventory decisions, determining the optimal trade-off between transportation and inventory costs. The au-thors compare two distribution strategies, namely direct truck delivery from a supplier to each customer, versus consolidated truck deliveries from a supplier to more than one customer. The optimal shipment size for each strategy is determined. Cetinkaya and Lee (2000) specify two important decisions to be made in shipment scheduling, namely the timing and the quantity of shipment, also known as temporal consoli-dation. The authors present an analytical model for coordinating in-ventory and transportation decisions in VMI systems. Tsao, Mangotra,
Lu, and Dong (2012) study an integrated facility location and inventory allocation problem for designing a distribution network with multiple distribution centers and retailers. The decisions to be made include the locations of distribution centers, the allocation of retail stores to the distribution centers and the inventory policies at the different locations in order to minimize total network cost. The decisions of facility location and inventory allocation are dependent on each other, since trans-portation cost influences the location decision and depends on the fre-quency of inventory replenishment. The authors develop nonlinear programming techniques to solve the optimization problems.
3. Problem description
Based on the literature review part presented in Section 2.1 a process map for handling of excavated rock and soil materials is developed, see
Fig. 1.
The process map in Fig. 1 together with the literature review pre-sented in Section 2.2 allow us to identify four questions that capture the interdependence between costs and environmental impact in rock and soil transport:
• how to enable reuse?
• how to decrease the number of transports?
• how to enable an efficient construction project without delays? • how to keep costs down?
One key issue answering these questions is the location of storage points of excavated rock and soil materials in relation to excavation points. However, reuse is not always performed at the same geograph-ical point as the excavation, which requires transports. Also, if reuse is done at the same location, there is a time lag between reuse and exca-vation and during this time the materials must be stored. Thus, the transport impact of a construction project depends on the available storage areas at different locations, the frequency and fill rate of the transports from excavation points to reuse areas as well as the routing between these locations. Furthermore, the cost depends on the number of machines and trucks needed as well as the number and size of storage locations.
Based on the above summary, the problem to be modelled can be formally described as follows: We have a number of excavation points with grinding machines, a number of intermediate storage points and a number of disposal points. The points are linked by a transport network. Vehicle fleets of different sizes and capacities are located at excavation
Fig. 1. Process map for handling of excavated soil and rock materials (based on
points and intermediate storage points and can be used to transport material within the whole transport network, but need to return to their depot (excavation point or intermediate storage point) each day. Over a predefined time horizon, material first has to be excavated and ground at excavation points and then transported to intermediate storage points. Intermediate storage points are used to hold the material while construction works are performed and supply the same material later to the excavation points for further construction works. Material that is not needed for construction works (excess material) is brought from
intermediate storage points to disposal points. An excavation point may need more or less material than originally excavated at that point. Due to working processes and space limitations at excavation points, an earliest date for the reuse of material is specified for each excavation point. The material reused by an excavation point is brought back from an inter-mediate storage point just in time. Excavation and construction at an excavation point are independent from each other in the sense that the excavation can be performed even if the demand of material is not ful-filled. However, only if a certain percentage of material has been
excavated and transported to an intermediate storage point, construc-tion material can be received at the excavaconstruc-tion point. Please note that demand at an excavation point can also be fulfilled directly with the material excavated at that point, without using intermediate storage points, to some extent. Disposal points are used to hold material which is not required for construction. However, even if located next to an excavation point, all intermediate storage points are independent and can be used to store material from and for any excavation point.
The aim of the present study is to generate detailed excavation and transportation plans for a project, as well as the required amount of intermediate storage points and the required storage area at each loca-tion. The objective function of the described problem encompasses two components: (1) Transport cost on the transport network are to be minimized, measured in terms of travel time of all vehicles over the whole time horizon. (2) Storage cost are to be minimized, measured as a fixed cost for storage locations depending on the size of the storage area, not the duration of storage. Several constraints need to be respected. Each excavation point possesses an earliest date for starting work and a due date for finishing work at that point. No more than a maximum amount of material can be excavated per day (due to limited resources for excavating and grinding). No more than a maximum amount of material can be brought back to excavation points per day. Vehicle ca-pacities and maximum working hours must not be exceeded.
4. Research design and solution approach
The present study follows a sequential exploratory mixed methods design (Puye & Hong, 2014). First a qualitative case study identifies the problem and the context of the problem. Second a quantitative optimi-zation method is used to evaluate possible scenarios of how to improve the problems identified in the case study. Mixed methods research in-tegrates qualitative and quantitative methods (Puye & Hong, 2014). According to Puye and Hong (2014) the combination of methods allows for a better understanding of new phenomena (qualitative methods) and the possibilities to measure their impact (quantitative methods).
4.1. Qualitative case methodology
An explorative study has the goal of reaching a more specific prob-lem description (Hellevik, 1984). A single case research design was selected because it allows for an in-depth understanding (cf. Eisenhardt, 1989; Flyvbjerg, 2006) and as the research is directed toward analysing a number of interdependent variables in complex structures (Dubois & Gadde, 2002). The single case selected, concerns the railway project Ostl¨anken in Sweden, with the focus on the part outside Norrk¨oping. To study this case represents a rare circumstance as well as it can be seen as a representative case of infrastructure projects (Yin, 2002). Data was collected through semi structured interviews with the project manage-ment (in total five interviews) and through documanage-ments such as drawings and plans including material to be excavated and material demand along the railway stretch. To increase understanding of the studied phenom-enon, a site visit was performed. Furthermore, two workshops were conducted to evaluate the developed optimization model and the sce-nario results. In the first workshop three persons participated except from the researchers and in the second 14 persons except from the researchers.
4.2. Optimization method
An optimization method based on Adaptive Large Neighborhood Search (ALNS) (Pisinger & Ropke, 2010) is implemented to treat the real-world problem investigated in the case study.
ALNS is a metaheuristic framework based on the idea of iteratively destroying and subsequently repairing solutions to explore large parts of the solution space and discover diverse local optima. Destroy operators remove certain parts of a solution according to some predetermined
selection strategy. The resulting partial solution is then reconstructed into a full solution with repair operators, commonly insertion heuristics. ALNS has been shown to perform well on a variety of problem classes, particularly when many constraints must be considered. ALNS is applied to investigate and evaluate several possible scenarios with respect to site layout and materials transport planning.
The interested reader is referred to a review by Pisinger and Ropke (2010) for a more detailed discussion of the related Large Neighborhood Search literature.
In our metaheuristic framework, 50 % of a solution is iteratively destroyed and subsequently repaired. Each ALNS run lasts at least 1 min and at most 10 min. After the first minute, we check if a certain number of consecutive iterations was carried out without success; if yes, then the run stops. We set this number as 1/3 of the total number of iterations already performed.
We embed the ALNS into a multi-start procedure, where we keep the selection probabilities of ALNS operators from one start to the next.
Martí, Resende, and Ribeiro (2013) state that combinatorial optimiza-tion problems require diversificaoptimiza-tion mechanisms to escape local optima during heuristic search procedures. One way to achieve diversification is to re-start the procedure from a new solution once a region has been explored. While in the present problem setting, defining neighborhoods to keep feasibility is hard, a new solution can relatively easily be con-structed, e.g. by varying the starting time of the different excavation points. In multi-start methods, each iteration (start) produces a solution (usually a local optimum) and the best overall is the algorithm’s output. We perform 50 multi-starts and keep the best run as the final solution.
Fig. 2 presents a flow chart of the ALNS algorithm applied to the case study.
An initial solution is generated before it is subsequently improved within the ALNS procedure. For each excavation point, a start date is randomly assigned such that the construction site could be finished on time, respecting the due date, with maximal workload per day. For each day of the planning horizon, the maximum amount of material that can be excavated is processed per day. The excavated material is brought to the nearest intermediate storage point. Please note that material that can directly be reused on site, is not considered. A random amount of ma-terial is brought back to an excavation point starting on the earliest date that material can be reused at that point. The daily amount of material brought from intermediate storage points to disposal points is randomly determined, considering material availabilities and requirements. Every day, each intermediate storage point with material inventory is randomly assigned to one of three sets. Set A contains intermediate storage points that do not provide any material on that day. Set B con-tains intermediate storage points that potentially provide material on that day to a disposal point. Set C contains intermediate storage points that potentially provide material on that day to an excavation point. Deliveries to disposal points or intermediate storage points are per-formed to the closest locations, while deliveries to excavation points are performed preferably from the nearest intermediate storage point that has enough material or, if there is none, from the nearest intermediate storage point that has any amount of material.
We explore three different destroy and repair mechanisms in the ALNS, which are selected by a roulette wheel method based on their success during the search. At the beginning, each mechanism has the same selection probability, which is increased if the particular operator leads to an improvement of the current solution. The operators are listed below, where mechanisms 1 and 2 can be combined flexibly, while destroy operator 3 is always complemented with repair operator 3.
Destroy Operators:
1) The first destroy operator removes deliveries to intermediate storage points, ranked by the maximum difference between the inventory peak and the second largest inventory value divided by the number of days the peak occurs. Deliveries are removed, such that the new peak is lower than the former second largest inventory value.
2) The second destroy operator removes used intermediate storage points with minimal days of occupation.
3) The third destroy operator removes trips from intermediate storage points to excavation points randomly.
Repair Operators:
1) The first repair operator reinserts material to an intermediate storage point with minimal distance to the delivering excavation point, such that no new material peak at the intermediate storage point is generated.
2) The second repair operator reinserts material to an intermediate storage point that is already used with minimal new material peak at that intermediate storage point.
3) The third repair operator reinserts material on another day in the future, on which material is already transported from the same in-termediate storage point to the same excavation point and can be combined on one trip (having sufficient spare loading capacity on the used vehicle).
All operators are randomized. Their rank is reverted and transferred into a selection probability.
5. The case of Ostl¨anken
The envisaged case is based on one of the largest infrastructure projects in Sweden, the Ostl¨anken railway construction, as illustrated in
Fig. 3. Today the site layout and materials planning process includes several steps. First, possible quantities to reuse of excavated soil and rock are identified. Second, the progress of the excavation is simulated with the help of a software called Magnet by Topcon. Third, based on the excavation and building progress simulation the need of number of storage locations and size of storage areas is calculated. This is also dependent on the planning horizon used in the project. Fourth, possible storage locations are identified and negotiations are started with land owners. In Sweden it is legally accepted to expropriate land, however it is not allowed to take more land than needed in possession and therefore one important aspect in this process is to identify the need of storage size and possible locations. The last two steps also have a relationship to the tendering of contractors and the geographical borders between con-tracts. Most contractors due to economical reasons do not benefit from moving soil and rock materials to another contractor within the same project and the most beneficial is to reuse the main part within the same contract. To summarize the above, the problem that the project man-agers deal with is to know what storage area size is needed, which among several locations to select and to determine the transport frequency.
5.1. Considered problem features
From the case description we extract four planning problem features: the size of storage area needed, the number of locations, the needed transport fleet and the geographical borders of the contracts. The geographical borders of the contracts will impact the possible planning horizon of the project. The transport frequency is dependent on the machinery and vehicles used in the project. We investigate these prob-lem features in different combinations to identify the impact on cost (storage, transport) and CO2 emissions.
1 Equal weight on objective function values (0.5 transport cost, 0.5 storage cost)
2 Priority weight on storage cost (0.2 transport cost, 0.8 storage cost) 3 Priority weight on transport cost (0.8 transport cost, 0.2 storage cost) 4 One-time planning, repeated day after day in a rolling horizon: All
excavation points start their work on the same day and process the maximal amount of material each day until all work has been finished. No planning over the whole time horizon is performed, but a greedy decision is taken every day. Starting with the excavation points processing the highest amount of excavated material, for each excavation point an intermediate storage point with minimal stored material is chosen, minimizing the distance between the excavation point and the intermediate storage point. Material is first brought from excavation points to intermediate storage points, before mate-rial is brought from intermediate storage points to excavation points or disposal points each day.
5 Whole-time planning for the total time horizon, each period con-sisting of one day
6 Homogeneous fleet of trucks with a load volume of 22 m3
7 Heterogeneous fleet of dumpers with a load volume of 13 m3
ca-pacity and trucks of 22 m3. Randomly, 50 % of locations are
equip-ped with dumpers and 50 % of locations are equipequip-ped with trucks.
5.2. Data
The data used to calculate and evaluate the different scenarios in the optimization model have been collected in the case study.
5.2.1. Vehicles
The vehicle types are trucks and dumpers, the specifications of these have been collected from Volvo Trucks website based on the advice of the planners given through interviews. Vehicles travel with a speed of 10–20 km/h. Working hours are limited to 8 h per day and loading and unloading times of vehicles vary between 2 and 3 min. Trucks have a weight of 3500 kg and can carry a capacity of 22 m3 of material.
Dumpers have a weight of 1500 kg and can carry a capacity of 13 m3 of
material. Excavation points are equipped with grinding machines, which
can maximally process a daily amount of 600 or 1200 m3 per excavation
point.
5.2.2. Storage and excavation locations
The potential locations of intermediate storage points have been identified from a zoomed-in version of Fig. 3 and with the help of in-terviews. A maximum number of potential locations were identified based on information from the project planner in the case. The locations of the excavation points have been given by the project planner in the case. The amount to be excavated over the whole time horizon takes a value between 4000 and 200,000 m3 at the different excavation points.
These have been provided from the planners that have calculated them with Magnet by Topcon. We assume that 1 m3 of rock and soil is
equivalent to 1500 kg. This assumption has been verified by the planners.
5.2.3. Costs
It was not possible to access actual costs in the case. Therefore, the storage cost of material per m3 was set to € 1 and the transport cost per
hour was set to € 1. CO2 equivalents are calculated with a formula based
on McKinnon and Pieczyk (2011). It considers vehicle weight (kg), amount of excavated material transported (m3), distance from the
ma-terial source to the mama-terial destination (m), distance from the mama-terial destination to the material source (m).
6. Experimental results
In Section 5.1 seven problem features were presented, that have through different combinations been transformed into six scenarios tested (Table 1). Computational results for different scenarios illustrate the implications of different problem features and policies. Table 1
summarizes the results of the different scenarios and enables a com-parison of total cost, transport cost, storage cost and CO2 emissions. The
scenarios are discussed in detail in the following subsections.
Table 1
Computational results for different scenarios illustrate the implications of different problem features and policies.
Scenario (features) Total Cost (€) weighted sum
Transport
Cost (€) Storage Cost (€) COEmissions (g) 2 1 50 % weight each on
transport and storage cost, homogeneous fleet (features: 1/5/ 6)
31.515 22.676 40.354 325.866.640
2 50 % weight on each transport and storage cost, heterogeneous fleet (features: 1/5/ 7) 34.703,5 30.145 39.262 368.476.324 3 20 % weight on transport cost, 80 % weight on storage cost, homogeneous fleet (features: 2/5/ 6) 35.309 28.757 36.947 464.154.993 4 80 % weight on transport cost, 20 % weight on storage cost, homogeneous fleet (features: 3/5/ 6) 26.562,2 20.986 48.867 288.252.413 5 15 intermediate storage points, homogeneous fleet, rolling horizon (features: 1/4/6) 58.193 20.560 95.826 304.074.161 6 5 intermediate storage points, homogeneous fleet, rolling horizon (features: 1/4/6) 57.393 28.862 85.924 468.683.113
6.1. Scenario 1: 50 % weight each on transport and storage cost, homogeneous fleet
The first scenario puts an equal emphasis on transport and storage cost, i.e. 50 % weight on each of the two components. All transports are performed with a homogeneous fleet of trucks with a cargo carrying capacity of 22 m3. The excavation plan for 23 excavation points is
displayed in Fig. 4 for Scenarios 1 and 2. The y-axis shows material excavation at each of the excavation points, while the x-axis contains the time horizon.
Fig. 5 shows the initial inventory plan for Scenario 1, before the optimization procedure, i.e. directly after applying the construction heuristic. On the y-axis, starting at the top, material stock at the two disposal points is illustrated, followed by the intermediate storage
Fig. 5. initial inventory plan (Scenario 1).
points.
After running the multi-start procedure with ALNS with an execution time of 60 min, the final inventory plan depicted in Fig. 6 is achieved. Compared to the initial plan, the number of intermediate storage facil-ities is reduced substantially. The percentage values displayed on the y- axis refer to the initial inventory at the respective intermediate storage points. In the initial solution, all potential intermediate storage facilities
are used without making use of spare capacity. Thus, initially storage cost is comparatively high. However, by perturbing the solution within the optimization procedure, the assignment of excavation material transports from excavation points to intermediate storage points, is improved to exploit synergies. Fig. 6 shows that storage space is effi-ciently used along the time horizon by reducing the number of inter-mediate storage points. In this way, storage cost is lowered by 60 % and
Fig. 7. final inventory plan (Scenario 2).
the total cost (transport and storage) is lowered by 50 %. The CO2
emissions of the final solution amount to 325.866.640 g.
6.2. Scenario 2: 50 % weight on each transport and storage cost, heterogeneous fleet
Fig. 7 shows the final inventory plan after running the multi-start
procedure with ALNS with an execution time of 85 min. On the y-axis, starting at the top, material stock at the two disposal points is illustrated, followed by the intermediate storage points. All excavation points and intermediate storage points are equipped either with a fleet of trucks with a cargo carrying capacity of 22 m3 or with a fleet of dumpers with a
cargo carrying capacity of 13 m3.
In the final solution of Scenario 2, most of the material is brought to
Fig. 9. initial inventory plan (Scenario 3).
one intermediate storage point, which is illustrated as number 48 on the y-axis of Fig. 7. Transport cost are 25 % higher compared to Scenario 1, where only homogenous trucks are used. This is because dumpers must go back and forth more often to carry the total amount of excavated material. The CO2 emissions of Scenario 2 amount to 368.476.324 g,
which is higher than in Scenario 1.
6.3. Scenario 3: 20 % weight on transport cost, 80 % weight on storage cost, homogeneous fleet
Fig. 8 shows the excavation plan for Scenario 3, where 20 % of emphasis is laid on transport cost, while 80 % is laid on storage cost. The y-axis shows material excavation at each of the excavation points, while the x-axis contains the time horizon. Figs. 9 and 10 show the initial
Fig. 11. excavation plan, 80 % transport, 20 % storage (Scenario 4).
inventory plan and the final inventory plan for this setting after a multi- start ALNS of 60 min computation time. On the y-axis, starting at the top, material stock at the two disposal points is illustrated, followed by the intermediate storage points.
As can be seen in Fig. 10, if most emphasis is laid on storage cost, only one intermediate storage point remains in the final solution. While storage cost is very low, transport cost is comparatively high since the
closeness of excavation points to their nearest intermediate storage point is not exploited. The CO2 emissions of this solution amount to
464.154.993 g.
Fig. 13. final inventory plan (Scenario 4).
6.4. Scenario 4: 80 % weight on transport cost, 20 % weight on storage cost, homogeneous fleet
Fig. 11 shows the excavation plan for Scenario 4, where 80 % of emphasis is laid on transport cost, while 20 % is laid on storage cost. The y-axis shows material excavation at each of the excavation points, while the x-axis contains the time horizon. Figs. 12 and 13 show the initial inventory plan and the final inventory plan for this setting after a multi-
start ALNS of 60 min computation time. On the y-axis, starting at the top, material stock at the two disposal points is illustrated, followed by the intermediate storage points.
Fig. 13 shows that several intermediate storage points are used in the final solution, if most emphasis is laid on transport cost. Transport cost are low compared to Scenario 3, where more importance is given to storage cost. In Scenario 4, the closeness of excavation points to their nearest intermediate storage point is exploited, thus leading to relatively
Fig. 15. inventory plan (Scenario 5).
low transport cost. The CO2 emissions of this solution amount to
288.252.413 g.
6.5. Scenario 5: 15 intermediate storage points, homogeneous fleet, rolling horizon
Fig. 14 shows the excavation plan determined on a rolling horizon for Scenarios 5 and 6. In this case, no future requirements are consid-ered. The y-axis shows material excavation at each of the excavation points, while the x-axis contains the time horizon. Each day, the maximum amount is excavated at each excavation point, all works starting at the same day.
Fig. 15 shows the inventory plan for Scenario 5, where material is stored at the closest intermediate storage point of each excavation point. On the y-axis, starting at the top, material stock at the two disposal points is illustrated, followed by the intermediate storage points. The CO2 emissions of this solution amount to 304.074.161 g.
6.6. Scenario 6: 5 intermediate storage points, homogeneous fleet, rolling horizon
Fig. 16 shows the inventory plan for Scenario 6, where only 5 in-termediate storage points are available. Material is stored at the closest intermediate storage point of each excavation point. On the y-axis, starting at the top, material stock at the two disposal points is illustrated, followed by the intermediate storage points. The CO2 emissions of this
solution amount to 468.683.113 g.
In Scenario 6, storage cost is lower than in Scenario 5, since the limited number of intermediate storage points enables to better occupy the available storage space.
7. Conclusions and recommendations
The purpose of this study was to develop a mixed method approach enabling scenario analysis of the relation between storage location and soil and rock material transport planning. First a case study was con-ducted to thoroughly understand the real-world problem features and the present working procedure, then a metaheuristic solution approach based on ALNS was developed to generate results for the real-world case. The problem features were combined into different scenarios that allowed us to explore the relationship between site layout planning, here with focus on inventory control decisions, and the transport planning of infrastructure development projects.
To improve transport efficiency in construction projects is an essential step in decreasing the environmental impact of the construc-tion industry. Transport efficiency can be improved through improved planning (Eriksson, 2019; McKinnon, 2018). Though, previous studies
within decreasing construction industry’s environmental impact have not focused on transport planning. Furthermore, sophisticated planning is uncommon and logistics is seen as a necessary evil in the management of rock and soil materials (Woodcock, 2015). Therefore, it is a contri-bution to explore how transport efficiency can be improved through linking the planning of site layout in infrastructure construction and rock and soil materials transport.
Thus this paper makes two main contributions:
1) With a mixed methods approach a way to decrease the environ-mental impact of infrastructure projects is presented, by improved planning of the transport and storage areas for rock and soil mate-rials. Problem features are investigated qualitatively, while the impact of action scenarios is evaluated with a quantitative optimi-zation approach.
2) A scenario analysis provides managerial implications of how to both decrease environmental impact of construction transport and improve construction productivity, i.e. considering a longterm planning horizon and aiming for high fill rates in vehicles that optimally use transport resources (e.g. High Capacity Transport or short sea shipping). The scenario analysis enables the developers and local authorities to understand their action space by visualizing the trade-offs, which are encountered in the planning processes. The literature review identifies a strong relationship between storage location and transports (e.g. Magnusson et al., 2019; Song et al., 2018), however the focus of these earlier studies has been the efficiency of the site or the ability to reuse materials, not the relationship between transport planning and storage locations. This study shows the impor-tance of considering the relationship between storage location/size and transports. Thus material transports from excavation points to inter-mediate storage points need to be considered when deciding on the transports from intermediate storage points back to excavation points and vice versa. The scenarios evaluated in this paper show that if focus is put on storage costs (i.e. few storage points), then transport costs and CO2 emissions will increase. Though, the risk of emphasizing the
transport cost and CO2 emissions is that many storage points will be
needed, which is practically not feasible. However, what is also shown is that with planning, i.e. overlooking the whole time horizon, there is a great possibility of decreasing both the need for storage points through increased utilization of the space and less transport costs and CO2
emission through optimal location within the project area. Thus, if site layout planning and transport planning is combined in strategic project planning, potential advantages of the interdependence between costs and CO2 emissions can be exploited.
This leads us into the managerial contribution of the paper. The issue of transport planning in construction cannot be left until the operational
stageof the project, which is the common way of acting today ( Wood-cock, 2015). It needs to be dealt with already in early planning stages, by the developers, to ensure available storage locations that allow to decrease the environmental impact of transports. These issues can be efficiently handled with the presented mixed methods approach, first analyzing the problem in detail, and then solving it with a metaheuristic solution approach as the algorithmic core of a decision support system. During a workshop the results of the model were compared to current practice in the studied case, and it was identified that the model com-plemented current practice, it can help developers’ procurement func-tion to understand how their decisions regarding contracts impact on the project outcome transport wise. Therefore, Fig. 17 presents a suggestion of a possible planning process showing the interdependence.
Further managerial contributions are related to specific findings of the scenarios tested.
•Comparing Scenario 1 (homogenous fleet) and Scenario 2 (heterog-enous fleet), we can confirm the statement of Akbarnezhad and Xiao (2017) that the volume of each transport renders the number of trips. Scenario 2 has higher total cost and CO2 emissions due to the smaller
size of the vehicles. This gives input to the results by Treiber and Bark (2016), which can be translated into policy impact, that to decrease the CO2 emissions from construction transport the present trend in
urban areas of limiting the truck size is not necessarily a way forward.
•Comparing Scenarios 5 and 6 (rolling horizon) with all other sce-narios, we see that these have the highest total cost and in the case of Scenario 6 (few storage points) also the highest CO2 emissions. Thus,
we show that the construction industry has to move away from the present way of planning construction transport on a daily basis (Woodcock, 2015) and instead consider long term planning of transports.
Declaration of Competing Interest
The authors report no declarations of interest.
Acknowledgement
We would like to acknowledge JPI Urban Europe that financed the MIMIC project which this research is part of. We would also like to thank the Swedish Traffic Administration for their time and support.
References
Akbarnezhad, A., & Xiao, J. (2017). Estimation and minimization of embodied carbon of buildings: A review. Buildings, 7(5), 1–24.
Behrends, S., Lindholm, M., & Woxenius, J. (2008). The impact of urban freight transport: A definition of sustainability from an actor’s perspective. Transportation
Planning and Technology, 31, 693–713.
Burns, L. D., Hall, R. W., Blumenfeld, D. E., & Daganzo, C. F. (1985). Distribution strategies that minimize transportation and inventory costs. Operations Research, 33 (3), 469–490.
Cetinkaya, S., & Lee, C.-Y. (2000). Stock replenishment and shipment scheduling for vendor-managed inventory systems. Management Science, 46(2), 217–232.
Chang, B., & Kendall, A. (2011). Life cycle greenhouse gas assessment of infrastructure construction for California’s high-speed rail system. Transport Research part D:
Transport and Environment, 16(6), 429–434.
Dablanc, L. (2007). Goods transport in large European cities: Difficult to organize, difficult to modernize. Transportation Research Part A: Policy and Practice, 41, 280–285.
Dalenstam, M. (2015). Att återanv¨anda schaktmassor. http://www.sverigesbergmateriali ndustri.se/images/pdf/handbocker_mallar_rad/rapport_masshantering.pdf.
Derrible, S. (2018). Urban engineering for sustainability. Cambridge, Massachusetts: The MIT Press.
Du, Q., Shao, L., Zhou, J., Huang, N., Bao, T., & Hao, C. (2019). Dynamics and scenarios of carbon emissions in China’s construction industry. Sustainable Cities and Society,
48.
Dubois, A., & Gadde, L.-E. (2002). Systematic combining: And abductive approach to case research. Journal of Business Research, 55, 553–560.
Eisenhardt, K. M. (1989). Building theories from case study research. The Academy of
Management Review, 14, 532–550.
Eras, J. J. C., Guiterrez, A. S., Capote, D. H., Hens, L., & Vandecasteele, C. (2013). Improving the environmental performance of an earthwork project using cleaner production strategies. Journal of Cleaner Production, 47, 368–376.
Eriksson, V. (2019). Transport efficiency: Analysing the transport service triad. Licentiate Thesis. Chalmers University of Technology.
Federgruen, A., & Zipkin, P. (1984). Computational issues in an infinite-horizon, multiechelon inventory model. Operations Research, 32(4), 818–836.
Flyvbjerg, B. (2006). Five misunderstandings about case-study research. Qualitative
Inquiry, 12, 219–245.
Hellevik, O. (1984). Forskningsmetoder i sociologi och statsvetenskap. Lund, Sweden: Natur och kultur.
Kaboli, A. S., & Carmichael, D. G. (2016). An examination of the DRET model and the influence of payload, haul grade and truck type on earthmoving emissions.
International Journal of Construction Management, 16(2), 95–108.
Magnusson, S., Lundberg, K., Svedberg, B., & Knutsson, S. (2015). Sustainable management of excavated soil and rock in urban areas – A literature review. Journal
of Cleaner Production, 93, 18–25.
Magnusson, S., Johansson, M., Frosth, S., & Lundberg, K. (2019). Coordinating soil and rock material in urban construction – Scenario analysis of material flows and greenhouse gas emissions. Journal of Cleaner Production, 241.
Martí, R., Resende, M. G. C., & Ribeiro, C. C. (2013). Multi-start methods for combinatorial optimization. European Journal of Operational Research, 226, 1–8.
Martinsen, U., & Huge-Brodin, M. (2014). Environmental practices as offerings and requirements on the logistics market. Logistics Research, 7.
McKinnon, A. (2018). Decarbonizing logistics - distributing goods in a low-carbon world. London: Kogan Page.
McKinnon, A., & Pieczyk, M. (2011). Measuring and managing CO2 emissions of European
chemical transports. Edinburgh, UK: Logistics Research Centre, Heriot-Watt
University.
Ozcan-Deniz, G., & Zhu, Y. (2017). Multi-objective optimization of greenhouse gas emissions in highway construction projects. Sustainable Cities and Society, 28, 162–171.
Pearce, A. R., & Ahn, Y. H. (2017). Sustainable buildings and infrastructure: Paths to the
future. New York, USA: Routledge.
Pisinger, D., & Ropke, S. (2010). Large neighborhood search. In M. Gendreau, & J. Y. Potvin (Eds.), Handbook of metaheuristics. International series in operations
research & management science (vol. 146, pp. 399–419). Boston: Springer.
Puye, P., & Hong, Q. N. (2014). Combining the power of stories and the power of numbers: Mixed methods research and mixed studies reviews. Annual Review of
Public Health, 35, 29–35.
Sanad, H. M., Ammar, M. A., & Ibrahim, M. E. (2008). Optimal construction site layout considering safety and environmental aspects. Journal of Construction Engineering and
Management, 134(7), 536–544.
Sandanayake, M., Luo, W., & Zhang, G. (2019). Direct and indirect impact assessment in off-site construction – A case study in China. Sustainable Cities and Society, 48.
Seo, M. S., Kim, T., Hong, G., & Kim, H. (2016). On site measurement of CO2 emissions during the construction phase of a building complex. Energies, 9, 599–612.
Smith, S. (1995). Productivity estimation of earthmoving operations using a discrete-event
simulation model. PhD Thesis. Edinburgh University.
Sobotka, A., & Czarnigowska, A. (2005). Analysis of supply system models for planning construction project logistics. Journal of Civil Engineering and Management, 11(1), 73–82.
Song, X., Xu, J., Shen, C., & Pena-Mora, F. (2018). Conflict resolution-motivated strategy towards integrated construction site layout and material logistics planning: A bi- stakeholder perspective. Automation in Construction, 87, 138–157.
Treiber, A., & Bark, P. (2016). Energieffektiva transporter av massgods i stora t¨atortsområden och storst¨ader. Rapport, 2016, 3. TFK, Sverige.
Tsao, Y.-C., Mangotra, D., Lu, J.-C., & Dong, M. (2012). A continuous approximation approach for the integrated facility-inventory allocation problem. European Journal
of Operational Research, 222(2), 216–228.
Woodcock, M. (2015). Construction logistics – Supply of bulk materials. In G. Lundesj¨o (Ed.), Supply chain management and logistics in construction – Delivering tomorrows built
environment (pp. 35–61). London: Kogan Page Limited.
Xu, J., Deng, Y., Shi, Y., & Huang, Y. (2020). A bi-level optimization approach for sustainable development and carbon emissions reduction towards construction materials industry: A case study from China. Sustainable Cities and Societies, 53.
Yin, R. K. (2002). Case study research - Design and methods. Thousand Oaks, California: Sage Publications Inc.
Aronsson, Håkan, & Huge-Brodin, Maria (2006). The environmental impact of changing logistics structures. International Journal of Logistics Management, 17(3), 394–415. B¨ackstrand, Jenny, & Fredriksson, Anna (2020). The role of supplier information
availability for construction supply chain performance. Production Planning & Control. https://doi.org/10.1080/09537287.2020.1837933
