Multicriteria decentralized decision making in logistic chains:a dynamic programming approach for collaborative forwarding of air cargo freight

摘要

Forwarding air freight cargo globally from the shipper's door to the door of the consignee is a complex logistic process and involves freight handling by several collaborating logistics companies. The door to door process is currently standardized by the International Air Transport Association with the industry initiative Cargo 2000. Many individuals with their own perspectives along the logistic chain decide how freight is transported and handled but have only limited insights how their decisions influence the follow-up decisions and the final on time delivery. A central planning authority can not be realized due to the heterogeneity of decision makers, the individual interests of the logistics companies, and their global operations. We argue that it is possible to overcome the deficient situation of complex transportation chains like in the air cargo industry by using optimization methods for decentralized decision making. This paper proposes a dynamic programming approach, which enables the decision makers in the decentralized situation to align their decisions better with the decisions of the involved partners. The approach guarantees that sensitive information of the logistics companies is kept local and only the most necessary information is shared along the logistic chain for a better planning. The transportation is planned with regard to multiple criteria, like the expected transportation costs and the probability to deliver the freight on time. We further show that our decentralized and multicriteria approach leads to better results compared to a local strategy that only exploits each decision maker's own perspective. Our approach is decentralized by nature and needs lean information exchange. Furthermore, it is as strong as a centralized approach that gathers all distributed information but that authorizes the logistic service providers to decide individually.