Abstract: Advances in performance of local area networks and microprocessors are providing significant computational capability at a relatively low cost. This paper is concerned with development of a distributed algorithm for minimum weight design of large structures on a network of workstations using biologically inspired genetic algorithms. Communication constructs from the software library Parallel Virtual Machine (PVM) have been used for message passing between the workstations. The algorithm has been applied to minimum weight design of two example structures. Performance estimates are provided based on the granularity and parallelization efficiency of the distributed model. The speedup of the distributed algorithm increases with the size of the structure, making it particularly suitable for optimization of large structures. For large structures, a high average speedup of about 10 is achieved using 11 workstations. The high scalability of the distributed genetic algorithm demonstrates that a cluster of workstations provides a cost‐effective alternative for high‐performance computing for coarse‐grained applications such as the GA‐based structural optimization.