A Multi Agent Based Organizational Architecture for Dynamic Pickup and Delivery Problem

摘要

Pickup and Delivery Problem (PDP) consists of searching an optimal set of vehicles and an optimal set of routes, one route by each vehicle, in order to pickup items from a set of origins and deliver them to another set of destinations. Pickup and delivery problem is a class of complex systems whose complexity is NP Hard. In PDP real life applications, heuristics and meta heuristics methods are used in order to obtain an acceptable solution in reasonable execution time. When unpredictable events, like for example path cut and vehicles failure, may occur during the PDP schedule execution, we say that the PDP is dynamic (DPDP) and in this case we have to revise this schedule. In this paper, we propose a multi agent architecture for DPDP based on an organizational architecture. Supported by a formal framework, the proposed architecture allows us to show, through a case study that computed solution for the studied problem could be done in a parallel manner which attenuates substantially the problem complexity.