A simple and effective algorithm for the MaxMin diversity problem
The challenge of maximizing the diversity of a collection of points arises in a variety of settings, including the setting of search methods for hard optimization problems. One version of this problem, called the Maximum Diversity Problem (MDP), produces a quadratic binary optimization problem subject to a cardinality constraint, and has been the subject of numerous studies. This study is focused on the Maximum Minimum Diversity Problem (MMDP) but we also introduce a new formulation using MDP as a secondary objective. We propose a fast local search based on separate add and drop operations and on simple tabu mechanisms. Compared to previous local search approaches, the complexity of searching for the best move at each iteration is reduced from quadratic to linear; only certain streamlining calculations might (rarely) require quadratic time per iteration. Furthermore, the strong tabu rules of the drop strategy ensure a powerful diversification capacity. Despite its simplicity, the approach proves superior to most of the more advanced methods from the literature, yielding optimally-proved solutions for many problems in a matter of seconds and even attaining a new lower bound.
KeywordsMaximum diversity MaxMin diversity Tabu search
- Andrade, M., Andrade, P., Martins, S., & Plastino, A. (2005). Lecture notes in computer science: Vol. 3503. GRASP with path-relinking for the maximum diversity problem (pp. 558–569). Google Scholar
- April, J., Glover, F., Kelly, J. P., & Laguna, M. (2003). Simulation-based optimization: practical introduction to simulation optimization. In Proceedings of the 35th conference on Winter simulation: driving innovation (pp. 71–78). Google Scholar
- Glover, F., & Laguna, M. (1997). Tabu search. Dordrecht: Kluwer Academic. Google Scholar
- Glover, F., Kuo, C., & Dhir, K. (1998). Heuristic algorithms for the maximum diversity problem. Journal of Information and Optimization Sciences, 19(1), 109–132. Google Scholar
- Silva, G. C., Ochi, L. S., & Martins, S. L. (2004). Lecture notes in computer science: Vol. 3059. Experimental comparison of greedy randomized adaptive search procedures for the maximum diversity problem (pp. 498–512). Google Scholar