A Reactive Multi-Agent Approach for Online (re)scheduling of Resources in Port Container Terminals

Abstract

The container transfer chain management should be carried out taking into consideration the maximum possible of environment interactions. For this reason, integrated approaches have to be investigated for solving scheduling problems in port container terminals. In our work, we propose a reactive multi-agent system for simultaneous (re)scheduling of vessel, quay crane, operator and trucks. The system contains a scheduling agent in the form of a heuristic whose performance is validated by comparing its results with an associated constraint programming model. The multi-agent system dedicated for embedded systems is tested with a reactive approach when the heuristic is able to reschedule on real time once a perturbation occurs. The robust solution obtained could also be used as a starting plan followed by rescheduling procedure for unexpected events in a proactive approach. Simulation study shows that the reactive approach provides less deviation between planned and actual schedules which guarantees the work smoothness and avoids flow instability.