Parallel genetic algorithm for multiknapsack problem
Qi Tang, Sun Ji Zhou, Guo Ji Chang · 2004
This paper begins by introducing the basic mechanics of genetic algorithm and discussing different ways to parallelize algorithm. A parallel genetic algorithm (PGA) is presented over a cluster of workstations by using the PVM library, which is used to handle communications among processors. Using the presented algorithm, the well-known 0-1 multiknapsack-problem is computed. Simulation results are presented to show how the performance of the PGA is affected by variations on the number of nodes, population size and migration interval. Results indicate that the performance of PGA on multiknapsack problem is sound and robust.