A New Powerful Global Optimizer over Continuous Spaces
Balancing the exploitation and exploration is an important consideration in the design of evolutionary algorithms (EAs). This package delivers a new swarm intelligence based optimizaiton algorithm with the name of natural aggregation optimization (NAA). NAA naturally mimics the self-aggregation behaviors of group-living animals to balance the exploitation and exploration in the search process. In NAA, the individuals are distributed into multiple sub-populations, and three main modules are designed to guide the search. Firstly, a built-in stochastic migration model is applied to balance the distributions of the individuals among different sub-populations and the unexplored space. Secondly, for the individuals in sub-populations, the located search module is applied to perform the exploitation. Thirdly, for the individuals who do not belong to any sub-population, the generalized search module is applied to do the exploration. Extensive experiments are performed to compare NAA with other state-of-the-art EAs. The results show that NAA has superior global searching capability over continuous spaces.
-
Fast
-
Reliable
You can install the algorithm simplly by Pypi
pip install NAAor clone the project from Github
git clone https://github.com/JasonGUTU/NAA.git-