Application of an Improved Diploid Genetic Algorithm for Optimizing Performance through Dynamic Load Balancing

The dynamic load balancing problem which can be defined as the effective redistribution of workload among the system processing units during execution time, is dynamic in nature where the load and the processing power of the system may change in time as units of work enter and leave the system and processing units are added to or removed from the processing pool. To address this problem, genetic algorithms are used in literature in different ways. In this study, promising results of applying an improved diploid genetic algorithm for load balancing a simulation of a network of processing units are reported. Key-words: Load balancing, changing environments, evolutionary optimizaton, genetic algorithms, diploid genetic algorithms.