Accelerated genetic algorithms: combined with local search techniques for fast and accurate global search

This paper discusses a combinatorial method to compute a global solution of a non-convex optimization problem. A hybrid algorithm is the synthesis of a genetic algorithm and a local search algorithm (e.g. nonlinear programming). The hybrid algorithm is not only faster than the genetic algorithm but also gives a more accurate solution. In addition, the length of chromosome required is much smaller. An abstract analysis of the hybrid algorithm is discussed based on which a convergent proof of this algorithm is derived. Finally, simulations are performed to show the efficiency of the algorithm