6533b858fe1ef96bd12b62fa

RESEARCH PRODUCT

Due Dates and RCPSP

subject

description

Due dates are an essential feature of real projects, but little effort has been made in studying the RCPSP with due dates in the activities. This paper tries to bridge this gap by studying two problems: the TardinessRCPSP, in which the objective is total tardiness minimization and the DeadlineRCPSP, in which the due dates are strict (deadlines) and the objective is makespan minimization. The first problem is NP-hard and the second is much harder, since finding a feasible solution is already NP-hard. This paper has three objectives: Firstly to compare the performance on both problems of well-known RCPSP heuristics - priority rules, sampling procedures and metaheuristics - with new versions we have developed that take due dates into consideration. Secondly, to present an instance generator that can generate instances with loose, medium, and tight due dates for computational study. And, finally, to adapt the technique of justification to deal with due dates and deadlines and to show its profitability.