No abstract

This paper studies an option contract for coordinating a supply chain comprising one risk-neutral supplier and two risk-averse retailers engaged in promotion competition in the selling season. For a given option contract, in decentralized case, each risk-averse retailer decides the optimal order quantity and the promotion policy by maximizing the conditional value-at-risk of profit. Based on the retailers’ decision, the supplier derives the optimal production policy by maximizing expected profit. In centralized case, the optimal decision of the supply chain system is obtained. Based on the decentralized and centralized decision, we find the coordination conditions of the supply chain system, which can optimize the supply chain system profit and make the profits of the supply chain members achieve Pareto optimum. As for the subchain, we also find the coordination conditions, which generalize the results of the supply chain with one supplier and one retailer. Our analysis and numerical experiments show that there exists a unique Nash equilibrium between two retailers, and the optimal order quantity of each retailer increases (decreases) with its own (competitor’s) promotion level.

A bi-attribute concave shortest path (BC-SP) problem seeks to find an optimal path in a bi-attribute network that minimizes a linear combination of two path costs, one of which is evaluated by a nondecreasing concave function. Due to the nonadditivity of its objective function, Bellman's principle of optimality does not hold. This paper proposes a parametric search method to solve the BC-SP problem, which only needs to solve a series of shortest path problems, i.e., the parameterized subproblems (PSPs). Several techniques are developed to reduce both the number of PSPs and the computation time for these PSPs. Specifically, we first identify two properties of the BC-SP problem to guide the parametric search using the gradient and concavity of its objective function. Based on the properties, a monotonic descent search (MDS) and an intersection point search (IPS) are proposed. Second, we design a speedup label correcting (LC) algorithm, which uses optimal solutions of previously solved PSPs to reduce the number of labeling operations for subsequent PSPs. The MDS, IPS and speedup LC techniques are embedded into a branch-and-bound based interval search to guarantee optimality. The performance of the proposed method is tested on the mean-standard deviation shortest path problem and the route choice problem with a quadratic disutility function. Experiments on both real transportation networks and grid networks show that the proposed method reduces the computation time of existing algorithms by one to two orders of magnitude.

The single machine scheduling problem which involves uncertain job due dates is one of the most important issues in the real make-to-order environment. To deal with the uncertainty, this paper establishes a robust optimization model by minimizing the maximum tardiness in the worst case scenario over all jobs. Unlike the traditional stochastic programming model which requires exact distributions, our model only needs the information of due date intervals. The worst case scenario for a given sequence that belongs to a set containing only n scenarios is proved, where n is the number of jobs. Then, the model is simplified and reformulated as an equivalent mixed 0-1 integer linear programming (MILP) problem. To solve the MILP problems efficiently, a heuristic approach is proposed based on a robust dominance rule. The experimental results show that the proposed method has the advantages of robustness and high calculating efficiency, and it is feasible for large-scale problems.

No abstract