Modified genetic algorithm with sampling techniques for chemical engineering optimization

In this work, we develop a new efficient technique to enhance the optimization ability, and to improve the convergence speed of genetic optimization algorithm. We investigate and introduce a number of sampling techniques to generate a good set of initial population that encourages the exploration through out the search space and hence achieves better discovery of possible global optimum in the solution space. The introduced sampling techniques include Latin hypercube sampling (LHS), Faure sequence sampling (FSS), and Hammersley sequence sampling (HSS). The performances of the proposed algorithms and a conventional genetic algorithm using uniformly random population are compared, both in terms of solution quality and speed of convergence. A number of test problems and a case study, optimization of multi-effect distillation, demonstrate the feasibility and effectiveness of the proposed techniques. With the same parameters, our technique provides a better solution and converge to the global optimum faster than the traditional genetic algorithm.