A GENETIC ALGORITHMIC FRAMEWORK FOR PROCESS DESIGN AND OPTIMIZATION

A general optimization framework for discrete and continuous problems based on genetic algorithmic techniques is presented. The proposed framework exhibits a structured information exchange leading to an efficient search procedure. The utility of the method is demonstrated by applying it to the design of an unsplit heat exchanger network. We also extend the basic idea of genetic algorithms to develop a new method, called the extended genetic search (EGS), for optimization in continuous spaces. The performance of the extended genetic search for several test cases of nonlinear unconstrained and constrained optimization problems is also presented. We also compare the performance of this technique with others, including simulated annealing. The proposed framework turns out to be quite robust and is able to locate global optimum solutions for problems where gradient-based algorithms fail. The results obtained were comparable to the ones obtained with simulated annealing. The implementational simplicity of the EGS framework makes it particularly useful as a tool for problems where approximate solutions are needed quickly.