A Large Neighborhood Search Heuristic for the Cumulative Scheduling Problem with Time-Dependent Resource Availability

Abstract : This paper addresses the cumulative scheduling problem with time-dependent total resource availability. In this problem, tasks are completed by accumulating enough amounts of resources. Our model is time-indexed due to the variation of the total resource availability. A Large neighborhood search (LNS) approach is developed. As an extension of our previous work, this paper contributes a new request removal method and an integration of a simulated annealing procedure to the LNS. Through the computational results, those modifications reduce the derived gap between the worst solutions and also the average derived gap.