A (1 + λ) evolutionary algorithm with the greedy agglomerative mutation for p-median problems

We consider the continuous p-median problem which is a popular mathematical model for both facility location and cluster analysis. For the p-median problems as well as very similar k-means problems, genetic and Variable Neighborhood Search algorithms with the greedy agglomerative heuristic procedure demonstrate their efficiency. However, the operation of such algorithms highly depends on the parameter of such procedure. We propose a new (1 + λ) evolutionary algorithm involving the greedy agglomerative procedure as the mutation operator which adjusts the parameter of this procedure during the operation of the algorithm. The advantage of the new algorithm was demonstrated by several examples including large-scale ones.