Multi-project scheduling with 2-stage decomposition

Official URL: http://192.168.1.20/record=b1306355 (Table of Contents)

A non-preemptive, zero time lag multi-project scheduling problem with multiple modes and limited renewable and nonrenewable resources is considered. A 2-stage decomposition approach is adopted to formulate the problem as a hierarchy of 0-1 mathematical programming models. At stage one, each project is reduced to a macro-activity with macro-modes, which are systematically generated by utilizing artificial budgets. The resulting single project network problem is a Multi-Mode Resource Constrained Project Scheduling Problem (MRCPSP) with positive cash flows. MRCPSP with positive cash flows is solved to maximize NPV and to determine the starting times and resource allocations for the projects. Using the starting times and resource profiles obtained in stage one each project is solved at stage two for minimum makespan. Three different time horizon setting methods, namely, relaxed greedy approach, artificial budget and Lagrangian relaxation are developed for setting the time horizon for MRCPSP with positive cash flows. A genetic algorithm approach is adopted to generate good solutions, which is also employed as a starting solution for the exact solution procedure. The result of the second stage is subjected to a post-processing procedure to distribute the resource capacities that have not been utilized earlier in the procedure. Since currently there are no data instances with the required structure, four new test problem sets are generated with 81, 84, 27 and 4 problems each. Three different configurations of solution procedures are tested employing the first three problem sets. A new heuristic decision rule designated here as Resource Return factor is presented and tested employing the fourth problem set.