An A-Team Based Framework for Logistics Scheduling

Abstract

Under the environment with complex sets of objectives and constraints, traditional approaches for logistics scheduling and planning typically result in a large monolithic model that is difficult to solve, understand, and maintain. The paper proposes a multi-agent constraint programming framework for logistics scheduling, especially under the dynamic and/or difficult circumstances such as traffic jam and natural and man-made disasters. In our framework, an asynchronous team of intelligent agents cooperate with each other to produce a set of non-dominated solutions that show the tradeoffs between objectives, and evolve a population of solutions towards a Pareto-optimal frontier. The framework has been successfully applied in real-world logistics scheduling, and demonstrate its capability to produce reliable and high-performance solutions with multi-objective optimization.