Matheuristics edition:2014 location:Hamburg date:11-13 June 2014
The Project Scheduling Problem consists of scheduling the processing times of jobs (or activities) contained in a project, while respecting precedence constraints between the jobs. This class of problems models many situations of practical interest in engineering and management science in general, and has been addressed by experts of various fields. The subject of this work is the Multi-Mode Resource-Constrained Multi-Project Scheduling Problem, an extension of the PSP that considers multiple projects with resource constraints and multiple possibilities to execute the activities. We produced a hybrid algorithm with several heuristics and Integer Programming (IP) based components: (i) a mode selection IP model; (ii) a project priority estimation heuristic; (iii) an IP constructive algorithm; and (iv) an IP local search algorithm. We present results of computational experiments on different PSP benchmark datasets to demonstrate the efficiency of the proposed approach. These experiments improved best known solutions for several instances of two different problem extensions.