Neuroevolutive Algorithms Applied for Modeling Some Biochemical Separation Processes

Abstract

Combining artificial neural networks with evolutive/bioinspired approaches is a technique that can solve a variety of issues regarding the topology determination and training for neural networks or for process optimization. In this chapter, the main mechanisms used in neuroevolution are discussed and some biochemical separation examples are given to underline the efficiency of these tools. For the current case studies (reactive extraction of folic acid and pertraction of vitamin C), the bioinspired metaheuristic included in the neuroevolutive procedures is represented by Differential Evolution, an algorithm that has shown a great potential to solve a variety of problems from multiple domains.

Keywords: Artificial neural networks; Differential evolution algorithm; Pertraction; Reactive extraction.