A new learning mechanism for resolving inconsistencies in using cooperative co-evolution model

Abstract

Many aspects of Software Engineering problems lend themselves to a coevolutionary model of optimization because software systems are complex and rich in potential population that could be productively coevolved. Most of these aspects can be coevolved to work better together in a cooperative manner. Compared with the simple and common used predator-prey co-evolution model, cooperative co-evolution model has more challenges that need to be addressed. One of these challenges is how to resolve the inconsistencies between two populations in order to make them work together with no conflict. In this position paper, we propose a new learning mechanism based on Baldwin effect, and introduce the learning genetic operators to address the inconsistency issues. A toy example in the field of automated architectural synthesis is provided to describe the use of our proposed approach. © 2014 Springer International Publishing Switzerland.