A parallel genetic algorithm for optimal designing of frame structure

A new evolutionary optimal designing method is presented for the designing of steel structures. We combine finite element structural analysis codes with genetic algorithm to minimize the total weight of the structure subject to external loads and constraints. The proposed algorithm is implemented using the Message Passing Interface(MPI) library over High Performance Computing Cluster-Dawning 4000. A master-slave paradigm is used to implement the proposed parallel distributed genetic algorithm. An example is given, and then, discuss the efficiency between the traditional sequential algorithm and the proposed parallel method.