A Genetic Algorithmic Approach for Solving the Multi-Resource General Assignment Problem: Design and Configuration Issues

摘要

This paper presents a viable GA based solution to the MR-GAP problem. Additional emperical testing and tuning is required. Preliminary results show that GA-MRG is competitive with other more traditional approaches in terms of solution quality and CPU utilization. It appears that the final solution appears generally within 20% of the generations required for convergence.. Additional efforts to leverage this may lead to further improvements in performance.rnGA-MRG was developed in the C language on a Gateway P5-60 PC using die Linux 1.3.70 operating system, die GCC 2.7.2 compiler and die emacs and GNU development environment.rnFurdier research in diis area will include tuning and hybridization of the GA with other techniques such as Tabu Search as well as evaluation of GA-MRG by solving practical business applications.