Parallel design methods for optimising lightweight structures

The gradient based approach to finding the optimum design of a structure is not naturally suitable for parallelisation by the 'divide and compute' str ategy, which simultaneously computes different independent designs, used by the random search based approach. This paper describes new workload distri bution strategies which are used with an existing fine grain parallel analy sis method in order to parallelise individual steps of the gradient based d esign procedure used in the authors' software VICONOPT. A Simple plane fram e program has been adapted to act as the testbed for all the strategies pre sented. Reasonably high levels of efficiency are reported for a relatively coarse grain Transtech Paramid parallel computer. (C) 2000 Civil-Comp Ltd a nd Elsevier Science Ltd. All rights reserved.