A combined genetic algorithm/simulated annealing algorithm for large scale system energy integration

A new algorithm named GA/SA (genetic algorithm/simulated annealing) is presented in this paper for solving a large scale system energy integration problem which is difficult to solve on the total process system level directly by traditional algorithm. The general GA has bean improved by using OCX (orthogonal crossover) and EC (effective crowding) operators, and the improved GA is combined effectively with an SA algorithm to avoid the common defect of early convergence. Numerical calculation results show that the new algorithm can converge faster than either SA or GA algorithms alone, and has much more probability of locating a global optimum. The convergence proof of the new algorithm is also given. GA/SA has been used to solve a 167 streams problem. A good result is achieved for improving the total process retrofit efficiency. (C) 2000 Elsevier Science Ltd. All rights reserved.