ABSTRACT
This chapter presents a scalable, distributed genetic algorithm for optimization of large structures on a network of workstations. The distributed genetic algorithm (GA) is organized in a master-slave configuration with one master process and a number of slave processes. In developing a distributed GA, a trade-off has to be made between the degree of parallelism and genetic search quality. Synchronous GA with centralized population produces exactly the same sequence of population generations as the sequential GA and, therefore, finds the same solution as the original GA. Distributed augmented Lagrangian GA requires some additional considerations. In this case, communication requirements vary with iterations and there is a larger amount of data transfer involved. Speedup is defined as the ratio of time taken to execute a task on processors to the time taken on a single processor with no communication or synchronization primitives inserted in it.
