Development of an algorithm for scheduling on independent parallel machines in a distributed context: Application to the management of preventive maintenance tasks

In this paper, an algorithm of tasks scheduling on different resources is proposed. We study especially the problem of complex system distributed across multiple sites. Each site is composed of a set of equipments to be maintained according to a predefined preventive maintenance policy. The proposed approach is consisting of two phases: the first ensures the assignment of the tasks to resources according to their availability; the second phase involves scheduling according to time constraints, due-dates and weightings of tasks. Two heuristics "LPT-HD" and "LPT-HD-EDD" are used to perform the algorithm with the objective to minimize the weighted sum of delays. The results are validated against results of ECT heuristic.