MULTI-PROJECT SCHEDULING WITH EXPLICIT LATENESS COSTS

Abstract

We propose a heuristic procedure for planning and scheduling multiple projects subject to limited resource availabilities. We depart from previous research in that explicit lateness costs for each project are considered. Our procedure involves aggregate analysis using linear programming to determine target resource loading profiles for each project that optimize trade-offs of lateness costs among projects, followed by detailed multi-project scheduling consistent with the target profiles. Target profiles and detailed schedules are iteratively modified through N iterations, where N is the number of projects. The procedure can be used to jointly schedule previously committed and newly proposed projects, as well as to assign due dates to proposed projects. We compare the performance of our procedure to that of the traditional minimum slack heuristic, as well to a simple extension of the minimum slack rule that accounts for lateness costs. On a set of 60 multi-project test problems adapted from the Patterson set of single-project problems, results are very favorable for our proposed algorithm.