The scheduling problem of serial multiproduct batch processes which attracts more and more attention in recent years is difficult to solve by some exact mathematical methods such as mixed integer programming (MIP) methods or branch and bound (BAB) when the size of the problem is large.This is mainly due to its NP complete nature.Simulated annealing (SA) has been proved to be efficient in finding an optimal solution for this kind of problem

but it requires considerable computing time for large problems.In this paper

an improved SA algorithm is proposed for solving the addressed scheduling problem for serial multiproduct batch processes with unlimited intermediate storage (UIS).The key to the improvement of the algorithm is the adoption of a non equilibrium and repeated annealing strategy

as well as various methods for generating neighboring solutions.Compared with presently available simulated annealing algorithms

the proposed algorithm gives better solution and runs faster in solving the addressed scheduling problem.Even for a very large scheduling problem

it can get reasonable result in acceptable computing time.The method appears to be promising for scheduling serial multi product batch processes.