A multi-agent approach for scheduling jobs and maintenance operations in the flowshop sequencing problem

Abstract

Many works refer to the scheduling problem of both preventive maintenance and production activities. Few works concern the dynamic scheduling problem of these two activities. This aspect is mainly concerned by corrective maintenance activities (equipment failure) and by condition-based maintenance (degradation). Decision-making tools are very helpful to schedule these activities. In this regard, the proposed multi-agent system provides a cooperation medium for scheduling independent jobs and maintenance operations in the flowshop sequencing problem based on the sequential and integrated strategies. The objective is then to optimise an objective function which takes into account both maintenance and production criterion. It also provides a framework in order to react to the disturbances occurring in the workshop. Moreover, it introduces a dialogue between two communities of agents (production and maintenance) leading to a high level of cooperation. A comparison of the solutions yielded by the strategies developed in this paper with the heuristic solutions given by Taillard (1993) is undertaken with respect to the minimisation of performance loss after maintenance insertion. The main point is to show how the proposed multi-agent system is used to generate a joint production and maintenance schedule that provides a better compromise between the satisfactions of respective objectives of these two functions.