RCPSP with Combined Precedence Relations and Resource Calendars

Abstract

This research introduces a new scheduling problem called the resource-constrained project scheduling problem with combined precedence relations and resource calendars (RCPSP/CPRs-cal). Combined precedence relations (CPRs) extend generalized precedence relations (GPRs) and can effectively represent all possible combinations of precedence relations between activities. Resource calendars reflect resource time constraints that are closer to real-life project scheduling. A new network diagram was designed for RCPSP/CPRs-cal; it uses nodes to unify activities and precedence relations. It solves a problem in which traditional networks are prone to form cycles when representing combined precedence relations. An extended genetic algorithm (GA-NND) is proposed for solving RCPSP/CPRs-cal efficiently; it uses the new network diagram for coding to generate the node scheduling scheme. A three-step guideline and an example are provided for schedulers in order to help them better use GA-NND to solve RCPSP/CPRs-cal. Computational experiments were carried out to show that the extended genetic algorithm can solve RCPSP/CPRs-cal more quickly and accurately than the traditional genetic algorithm. The new problem contributes to developing a practical decision support system for resolving real-life constraints in projects.