Adaptive Simulated Annealing for Global Optimization in LS-OPT

Summary: The efficient search of global optimal solutions is an important contemporary subject. Different optimization methods tackle the search in different ways. The gradient based methods are among the fastest optimization methods but the final optimal solution depends on the starting point. The global search using these methods is carried out by providing many starting points. Other optimization methods like evolutionary algorithms that mimic the natural processes like evolution, and simulated annealing that emulates the metal cooling process via annealing can find the global optima but are criticized due to high computational expense. The adaptive simulated annealing algorithm has been proposed to be an efficient global optimizer. This algorithm is implemented in LS-OPT. A few analytical examples and meta-model based engineering optimization examples are used to demonstrate the efficiency of the global optimization using ASA. The optimization results are also compared with the existing LFOPC and genetic algorithm optimization methods.