Cybernetic optimization by simulated annealing: Accelerating convergence by parallel processing and probabilistic feedback control (Q1922638)

The new concept of using a feedback control mechanism based on an error signal which comes from many parallel computers for solving optimization problems is presented. The convergence of the simulated annealing algorithm is accelerated by a probabilistic feedback control scheme. This scheme uses many parallel processors to solve the same or related combinatorial optimization problems, and they are coupled by a probabilistic measure of quality which is further used to generate an error signal in feedback control. These ideas were tested experimentally to solve the independent set and vertex cover problems.