Genetic Annealing Optimization: Design and Real World Applications

Both simulated annealing (SA) and the genetic algorithms (GA) are stochastic and derivative-free optimization technique. SA operates on one solution at a time, while the GA maintains a large population of solutions, which are optimized simultaneously. Thus, the genetic algorithm takes advantage of the experience gained in the past exploration of the solution space. Since SA operates on one solution at a time, it has very little history to use in learning from past trials. SA has the ability to escape from any local point; even it is a global optimization technique. On the other hand, there is no guarantee that the GA algorithm will succeeded in escaping from any local minima, thus it makes sense to hybridize the genetic algorithm and the simulated annealing technique. In this paper, a novel genetically annealed algorithm is proposed and is tested against multidimensional and highly nonlinear cases; fed-batch fermentor for Penicillin production, and isothermal continuous stirred tank reactor CSTR. It is evident from the results that the proposed algorithm gives good performance.