Are computers better for optimizing structures than design engineers?

Thursday, 1 March, 2012

The construction industry is cost driven, yet at the same time it searches for better quality, more esthetics and higher functionality. Combining these apparently contradictory goals is well-known in the world of economics. The discipline Operations Research deals with the application of advanced analytical methods to help making better decisions. Operations Research is often concerned with determining the maximum (of profit, performance, or yield) or minimum (of loss, risk, or cost) of some real-world objectives. Originating in military efforts before World War II, its techniques have grown to concern problems in a variety of industries.

So did Scia when starting a research project together with the University of Prague in 2006. From this research project a new algorithm could be developed which is now available in Scia Engineer, as a module named "Optimizer". A Scia Engineer user specifies "what to optimize" (e.g. weight, shape, ...) and chooses the strategy (with constraints). The Optimizer iterates automatically towards the optimum and stores all steps in a table with the best solution highlighted.

Several algorithms from Operations Research were implemented, to name a few: gradient method, Nelder-Mead heuristic method, differential evolution method and others.

And yes, it works. Practical design examples have been worked out (e.g. choice of best post-tensioned cable tendons in a bridge) with amazing cost reductions (> 15%) compared with the designer's proposal.

For those interested in more details, please download our white paper or contact Scia for a technical discussion.