Parallel algorithm of simulated annealing with modified generation mechanism

The authors propose a general parallel algorithm for simulated annealing with a modified generation mechanism The proposed algorithm has the following advantages: (i) it can be generally applied to various kinds of combinatorial optimization problems, (ii) it is easily implemented as a computer program, (iii) it is less sensitive to the value of the control parameter than the standard algorithm, and (iv) it can generate better solutions than the standard algorithm. The first and second advantages are clarified from the structure of the proposed algorithm. The third and the fourth advantages are shown from simulation results for a flow shop scheduling problem.<<ETX>>