Process optimization via simulated annealing: Application to network design

W. B. Dolan, P. T. Cummings, M. D. LeVan
1989 AIChE Journal  
Simulated annealing is a multivariable optimization technique based on the Monte Carlo method used in statistical mechanical studies of condensed systems and follows by drawing an analogy between energy minimization in physical systems and costs minimization in design applications. In this paper, simulated annealing is introduced and reviewed. The utility of the method for optimization of chemical processes is illustrated by applying it to the design of pressure relief header networks and heat
more » ... xchanger networks. Laarhoven (1985) report a simulated annealing method for the N-city traveling salesman problem that yields solutions within 2% of the global optimum and grows as N3. Algorithms which are guaranteed to find the global optimum for such problems increase in computational time much more quickly and, in the worst case, are exponential in N. With simulated annealing, the minimization of a global objective function is approached directly. When applied to a problem of cost minimization, moves are accepted and rejected based on a cost function. No heuristic arguments are needed.
doi:10.1002/aic.690350504 fatcat:2asr5xzuwzb4nelehv4towfodq