Differential Evolution with Tabu List for Solving Nonlinear and Mixed-Integer Nonlinear Programming Problems

Differential evolution (DE), a population-based direct-search algorithm, has been gaining popularity in the
recent past due to its simplicity and ability to handle nonlinear, nondifferentiable, and nonconvex functions.
In this study, a method, namely, differential evolution with tabu list (DETL), is described and evaluated for
solving constrained optimization problems encountered in chemical engineering. It incorporates the concept
of tabu search (TS) (i.e., avoiding revisits during the search) in DE mainly to improve its computational
efficiency. DETL is initially applied to many nonlinear programming problems (NLPs) involving 2−13
variables and up to 38 constraints. It is then tested on several mixed-integer nonlinear programming problems
(MINLPs) encountered in chemical engineering practice. The performance results of DETL, DE, and modified
differential evolution (MDE) (Babu, K. V.; Angira, R. Comput. Chem. Eng. 2006, 30, 989), for both NLPs
and MINLPs, are presented, and the relative performance of the three methods is discussed.