A New Approach for Resource-Constrained Multi-Project Scheduling

Construction and facilities maintenance projects involve a large number of people and tasks with resource constraints and precedence constraints. This paper presents a new approach to model this problem as a resource-constrained multi-project scheduling problem (RCMPSP) with cost minimization. The scheduling problem is first decomposed into two sub-problems: schedule generation and sequencing. For the schedule generation problem, an effective forward and reverse schedule generation (FRSG) method is developed to generate a feasible solution for a given valid sequence. For the sequencing problem, a novel complete local search with memory approach embedded with FRSG is proposed to find the solution which has the best objective value. The proposed approach has been tested on the benchmark instances. Computational results show that it performs very well in terms of both effectiveness and efficiency.