Multi-modal Optimization Algorithms

One of the niches of evolutionary algorithms is their ability to solve multi-modal problems. For these problems, niching based EC methods were able to find and maintain multiple optimal solutions simultaneously in a single simulation. At COIN, we are working on different ways of finding multiple optimal solutions. The effort is spent in developing a parameter-less approach. Some of past efforts (Deb and Saha, 2000) were spent in devising a multi-objective based method and suggestion of test problems having scalable number of optima. The work has also suggested, for the first time, constraint multi-modal problems having scalable number of optima.