Optimization of a Process Synthesis Superstructure Using an Ant Colony Algorithm

The optimization of chemical syntheses based on superstructure modeling is a perfect way for achieving the optimal plant design. However, the combinatorial optimization problem arising from this method is very difficult to solve, particularly for the entire plant. Relevant literature has focused on the use of mathematical programming approaches. Some research has also been conducted based on meta‐heuristic algorithms. In this paper, two approaches are presented to optimize process synthesis superstructure. Firstly, mathematical formulation of a superstructure model is presented. Then, an ant colony algorithm is proposed for solving this nonlinear combinatorial problem. In order to ensure that all the constraints are satisfied, an adaptive, feasible bound for each variable is defined to limit the search space. Adaptation of these bounds is executed by the suggested bound updating rule. Finally, the capability of the proposed algorithm is compared with the conventional Branch and Bound method by a case study. Two approaches are presented to optimize process synthesis superstructure. Firstly, a mathematical formulation of a superstructure model is presented. Then, an ant colony algorithm with feasible bounds for each variable is proposed for solving this nonlinear combinatorial problem. The capability of the proposed algorithm is compared with the conventional Branch and Bound (B & B) method by a case study.