Self-Configuring (1 + 1)-Evolutionary Algorithm for the Continuous p-Median Problem with Agglomerative Mutation

The continuous p-median problem (CPMP) is one of the most popular and widely used models in location theory that minimizes sum distances from known demand points to sought called centers or medians. This NP-hard also useful for clustering (automatic grouping). In this case, are considered as cluster centers. Unlike similar k-means model, less sensitive noisy data appearance outliers (separately located do not belong any cluster). Local search algorithms including Variable Neighborhood Search well evolutionary demonstrate rather precise results. Various based on use greedy agglomerative procedures capable obtaining very accurate results difficult improve with other methods. computational complexity such limits their large problems, although computations massively parallel systems significantly expand capabilities. addition, efficiency highly dependent setting parameters. For majority practically important can choose a efficient algorithm procedures. However, parameters algorithms, which ensure high efficiency, predict. We introduce concept AGGLr neighborhood application procedure, investigate depending its parameter r. Using similarities between local (1 + 1)-evolutionary ability latter adapt parameters, we propose new procedure automatically tuned Our does require preliminary tuning r adjusting online, thus representing more versatile tool. advantages shown experimentally problems volume up 2,000,000 points.