Location routing problems (LRPs) involve determining facility locations and vehicle routes. We extend the LRP model from the viewpoint of real-world delivery. First, we extended the capacity constraint of the vehicle. Previous studies considered only one type of capacity constraint, such as quantity. However, in the real world, there are at least two more types of capacity constraints: volume and weight. Therefore, we added multiple attributes to orders and vehicles so that multiple capacity constraints are applied. Next, we extended the delivery constraint. In a previous work, it was assumed that products are delivered to each customer only once. This means that every order must be equal to or smaller than the vehicle's capacity. If there is only one type of capacity constraint in the model, orders might be divided due to the available capacity. However, our model has multiple attribute constraints, so it is hard to divide the orders so as to minimize logistics cost. In our model, the orders are divided optimally. Our model decides not only the delivery route of the vehicle, but also the delivery volume and weight for each customer. The model was formulated as a mixed-integer programming problem. Numerical experiments show that our model can divide orders so as to minimize logistics cost.
During contract negotiations between two companies in a supply chain, the total profit of the two companies based on the agreed contract price and contract volume is usually lower than the monopoly profit due to the asymmetry of the information. Additionally, with such an agreement, there is a large difference in the profit distribution between the two companies. The effectiveness of preventing profit and profit sharing together with information sharing between companies has been confirmed in many previous studies, but it is very difficult to realize sharing of confidential information such as cost information among non-affiliated companies with low reliability. In this study, by proposing a collaborative negotiation method with a coordinator, prediction function and negotiations guide function, cost information sharing could be realized in a pseudo-manner. Mitigation effects for declining SC profit and profit allocation are shown in the results of this study.
To reduce costly order picking activities, many companies configure their warehouse with a forward area and a reserve area. The former is a small area where the most popular parts can be conveniently picked. The latter is used for replenishing the forward area and storing parts that are not assigned to the forward area. As pickers consume the inventory in the racks, there is a risk of stockout. Having sufficient inventories in the forward area of a warehouse is essential for warehouse operations. Parts are replenished using a forklift because of their heaviness. Multiple parts can be replenished at the same time, but depending on the combination of parts to be replenished, unnecessary working time may occur. Reducing the number of parts stored in the forward area enables a more compact forward area, thus reducing picking effort there, but requires more frequent replenishment from the reserve area to the forward area. Based on real data, this paper studies setting the reorder level and determining the combinations of parts to be replenished in order to reduce working time in the warehouse of an automotive factory. We model this problem as an integer programming problem. Furthermore, we present an algorithm to estimate working time quickly, even in the case of a large-scale problem.