Application of artificial life approaches to optimization

In this study, several methods of artificial life (AL) are introduced, and their characteristics and applicability to optimization problems are discussed. As AL methods, we consider the genetic algorithm (GA), the Hopfield neural network model, the Boltzmann machine (BM)neural network model, the randomized algorithm (RA), and the neural network model using the self-organizing map (SOM). These;methods are used to solve the Nondeterministic polynomial (NP) complete problem and the optimization problem with discretized parameters. Through a numerical example, the effectiveness of the RA and GA models is demonstrated. Finally, an effective method, which can be applied to civil engineering problems, is also proposed based on the RA model. Its effectiveness is demonstrated through several examples including the reliability-based optimization of a girder bridge.