Temperature Optimization in Non-isothermal Tubular Reactor using Genetic Algorithm

Abstract

Genetic algorithm (GA) is a heuristic search algorithm that is inspired by evolution. It is a powerful optimization tool that uses the stochastic procedure with populations of initial guesses rather than using a single value like gradient-based methods. This prevents GA from being trapped in a local optimum. In the present work, GA applications to industrial optimization problems are thoroughly reviewed to get a perspective on different variations of genetic algorithms being used in industries. Subsequently, GA is applied to an industrial tubular reactor system where the technique is used to determine the optimum feed temperature at reactor inlet so that the product attains desirable temperature at the reactor outlet. In addition to successful application of GA, some other performances such as effect of mutation function and selection technique on the number of iterations are also investigated.