An evolutionary strategy for fed-batch bioreactor optimization; concepts and performance

An evolutionary program, based on a real-code genetic algorithm (GPI), is a pplied to calculate optimal control policies for bioreactors. The GA is use d as a nonlinear optimizer in combination with simulation software and cons traint handling procedures. A new class of GA-operators is introduced to ob tain smooth control trajectories, which leads also to a drastic reduction i n computational load. The proposed method is easy to understand and has no restrictions on the model type and structure. The performance and optimal t rajectories obtained by the extended GA are compared with those calculated with two common methods: (i) dynamic programming, and (ii) a Hamiltonian ba sed gradient algorithm. The GA proved to be a good and often superior alter native for solving optimal control problems. (C) 1999 Elsevier Science B.V. All rights reserved.