Parametrized GRASP Heuristics for Three-Index Assignment
Constructive greedy heuristics are algorithms that try to iteratively construct feasible solutions for combinatorial optimization problems from the scratch. For this they make use of a greedy scoring function, which evaluates the myopic impact of each possible element with respect to the solution under construction. Although fast, effective, and even exact for some problem classes, greedy heuristics might construct poor solution when applied to difficult (NP-hard) problems. To avoid such pitfalls we suggest the approach of parametrizing the scoring function by including several different myopic aspects at once, which are weighted against each other. This so-called pgreedy approach can be embedded into the metaheuristic concept of GRASP. The hybrid metaheuristic of GRASP with a parametrized scoring function is called parametrized GRASP heuristic (PGRASP). We present a PGRASP algorithm for the axial three index assignment problem (AP3) and computational results comparing PGRASP with the classical GRASP strategy.
KeywordsFeasible Solution Greedy Algorithm Combinatorial Optimization Problem Construction Phase Candidate Point
Unable to display preview. Download preview PDF.
- 1.Aiex, R.M., Resende, M.G.C., Pardalos, P.M., Toraldo, G.: GRASP with Path Relinking for Three-Index Assignment. INFORMS Journal on Computing 17(2), 224–247 (2005), Software available at URL http://www.research.att.com/~mgcr/
- 6.Fügenschuh, A.: The Integrated Optimization of School Starting Times and Public Transport. PhD Thesis, Logos Verlag Berlin, 165 pages (2005)Google Scholar
- 8.Glover, F.: Tabu Search and Adaptive Memory Programming - Advances, Applications and Challenges. In: Barr, R.S., Helgason, R.V., Kennington, J.L. (eds.) Interfaces in Computer Science and Operations Research, pp. 1–75. Kluwer Academic Publishers, Dordrecht (1996)Google Scholar
- 9.ILOG CPLEX, ILOG, Inc., 1080 Linda Vista Ave. Mountain View, CA 94043. Information available at URL http://www.ilog.com/products/cplex/
- 16.Resende, M.G.C.: Test Data for the Three Index Assignment Problem, online available at URL http://www.research.att.com/~mgcr/data/