Parallel simulated annealing using simplex method

An improved edition of hybrid algorithms combining simulated annealing (SA) with and simplex method (SM), a parallel simulated annealing using simplex method (PSASM), was proposed, and applied to structural optimization. The simplex can be regarded as a population including (n+1) individuals for n-diemsional optimization problem, The excellent individuals of the simplex can be spread out using reflection, expansion or contraction operation. The PSAMS was evaluated in two functional problems, which included a two-dimensional function and the Golinski's speed reducer problem. The results of the PSASM were found to be able to locate the near-optimal solution at probability 100%, with low computational cost. The PSAMS was also applied to a 10-bar truss structural design problem, and was found competitive when compared with other stochastic optimization algorithms. More implementations for the PSASM further examined are expected to improve its performance.