Road transportation generates significant costs for firms that deliver and collect products. Next to that, it results in emissions. Food supply chains cause additional emissions because of the extra energy needed to guarantee product quality. Moreover, current transportation systems are inefficient since the available capacity is not optimally used. Cooperation between food supply chain actors could provide opportunities to reduce costs and emissions and improve eco-efficiency, which is defined as solutions for which it is impossible to improve the environmental objective without worsening the economic objective. In food supply chains, it is also important to guarantee food quality. However, reducing costs, emissions, and quality decay do not always go hand in hand, and trade-offs need to be made. Moreover, despite the benefits, companies hesitate to implement cooperation because it might bring advantages to competitors and they find it difficult to agree on gain sharing. To find out how cooperation can improve eco-efficiency in food logistics, we need decision support models that can capture these complexities. Therefore, the main research question that this thesis aims to answer is: Which decision support models can be used to design eco-efficient logistics cooperation in food supply chains?
All studies in this PhD thesis are based on a case study on a Dutch retail cooperative, where several smaller retail organisations cooperate by forming a buying organisation. By jointly purchasing their products, they can negotiate a lower unit price. The retailers currently hire different logistics service providers to pick up their orders from their shared distribution centre and bring them to their own distribution centre. From there, they distribute the products to their supermarket outlets. Currently, the retailers arrange their logistics individually, but they are considering to cooperate to reduce costs, emissions and quality decay.
In this thesis, the effects of different forms of logistics cooperation between food supply chain actors are analysed using existing optimisation models, which we extended to account for temperature control and food quality. Using these extended models, routing and inventory are optimised to minimise costs, emissions, and quality loss. Moreover, trade-offs between the objectives are established. Also, we proposed a method to divide cooperative gains, not only based on costs but also on emissions. This way, eco-efficient forms of logistics cooperation are rewarded and stimulated.
In Chapter 2, we extend the Load Dependent Vehicle Routing Model such that it accounts for the extra costs and emissions related to temperature control. We show that temperature control can significantly affect costs and emissions and thus the optimal routing. This extended model can be used to test the effect of new cooling technologies on the costs and emissions of routing.
In Chapter 3, we introduce the Quality Driven Vehicle Routing Problem. This problem is modelled and used to more realistically quantify how food degrades during distribution processes. We consider effects of outside temperature, door openings, and differences in optimal temperatures for different products. When transporting temperature-sensitive products, minimising quality loss results in multiple routes with less stops per route whereas minimising costs or emissions results in longer routes. The negative quality effects of multi-stop routes can be mitigated by adjusting driving speed, unloading rate, cooling rate, and by setting a quality threshold level.
In Chapter 4, we compare the effect of different forms of cooperation in temperature-controlled transportation on cost and emissions. Joint route planning (JRP, in which daily transport decisions are optimised cooperatively) is compared to vendor-managed inventory (VMI, in which multi-day routing and inventory decisions are optimised cooperatively) and to a non-cooperative scenario using vehicle routing and inventory routing models. In JRP, there is one optimal solution for minimising costs and for minimising emissions. For VMI however, additional savings in both objectives are obtained but there is a set of alternative eco-efficient solutions and partners need to choose which of those solutions (i.e. cooperative routing and inventory plans) they prefer. Also, in VMI there exists a trade-off between product age and emissions: less frequent inventory replenishment leads to reduced emissions but it also to a higher average product age.
In Chapter 5, we study how the monetary benefits of VMI can be allocated. We discuss that gain allocations should reflect both contributions to savings in costs and emissions. That way, gain allocation can be used to stimulate eco-efficient forms of cooperation. A green IRP model is used to quantify cooperative benefits and establish all possible eco-efficient cooperative solutions. For each solution, we allocate monetary benefits based on costs and emissions using the Shapley value. This approach results in cost savings for all partners that help reducing impacts.
In this thesis, we adjusted VRP and IRP models to account for temperature control and perishability. Using these OR models, we found that food logistics cooperation can result in significant economic and environmental benefits. The findings of all studies can be summarised in three main concluding statements: (i) temperature control influences costs and emissions of cooperative routes, and cooperative routing influences food quality. Therefore, these food specific aspects should be considered in cooperative logistics; (ii) dependent on the intensity of the cooperation, it can result either in one optimal solution for both costs and emissions, or a set of eco-efficient solutions; and (iii) to stimulate forms of cooperation that both reduce costs and emissions, we should allocate cooperative gains based on partner’s contributions to both indicators.