A Flexible Solution Framework for Optimal Decision in Logistics Systems

This paper concentrates on the methodology of using evolutionary computation (EC) for decision optimization problems in the context of logistics systems. We introduce a new perspective for practitioners to synthesize the problem-specific optimizer. A new EC model is proposed for effectively use of the evolutionary search framework for practical logistics applications. The general principle of "trade-off between exploration and exploitation" is significantly operable under our proposed solution framework with decoupled exploration and exploitation operators. Furthermore, the typical requirements of different logistics problems are discussed and the guidelines for deign and analysis of the tailored algorithms, derived from our generic solution framework, are comprehensively provided based on a series of analytical results. As an example, a typical warehousing task in the rotary rack Storage/Retrieval system is studied to demonstrate the implementation details for practical problems. Additionally, comparison experiments are carried out and related results illustrate the benefits of proposed solution framework and design methodology for practical optimization problem.