ABSTRACT
Multimodal cases con taining many local opti ma such as shown in Figure 7.1 (Al-Mashikh and Nakai, 1987) are quite infrequent to run into in food research. As a result, the response surface methodology (RSM) and optimization based on it searching for the maximum or the minimum on smooth surfaces is rather popular (Banga et al., 2003). However, the advantage of stochastic approach in global optimization was extensively studied so that there is no need of a priori knowledge on the mechanism of the optimization projects in concern. The number of papers published on the global optimization in chemistry and engineering had dramatically increased since 1900. This recent trend in publications for global optimization may be generated due to the introduction of a new algorithm, i.e., ‘‘genetic algorithm’’ that was a general methodology used to search for a solution space in a manner analogous to the natural selection procedure in the biological evolution (Holland, 1975). This optimization framework is able to provide the global optimum when
conventional gradient-based algorithms have failed. The results of computations thus obtained were comparable to that derived from more recent ‘‘simulated annealing’’ (Androulakis and Venkatasubramanian, 1991).
