Temperature Parallel Simulated Annealing with Adaptive Neighborhood for Continuous Optimization Problem

In this study, a temperature parallel simulated annealing with adaptive neighborhood (TPSA/AN) for continuous optimization problems is introduced. TPSA/AN is based on the temperature parallel simulated annealing (TPSA), which is suitable for parallel processing, and the SA that Corana developed for continuous optimization problems. The moves in TPSA/AN are adjusted to have equal acceptance rates. Because of this mechanism, the proposed method provides global search in the processors of parallel computers for high temperatures and local search in the processors for low temperatures in TPSA/AN. Therefore, all the processors are used for searching very efficiently. The TPSA/AN is evaluated for the standard test functions, and it is found that adopting the adaptive neighborhood range increases the searching ability of TPSA.