Exponential evolutionary programming without self-adaptive strategy parameter

Evolutionary programming (EP) uses strategy parameter with self-adaptation. This strategy parameter corresponds to a search step size in solution search algorithm. Exponential evolutionary programming (EEP) uses exponential mutation instead of Gaussian mutation of conventional evolutionary programming (CEP). Therefore, the search step size of EEP depends on the parameter value of exponential distribution as well as self-adaptation. Generally, the strategy parameter has to decrease its value with evolution progress. For the sake of this purpose, the parameter value of EEP has to augment the self-adaptation of EP. However, it is not so easy to find the fine tuning parameter value of EEP with linkage to the self-adaptation in actual computation. Considering these situations, we propose here new EEP (nsEEP) without self-adaptive strategy parameter. Instead of self-adaptation, the parameter value of EEP changes automatically with evolution progress. In this paper, we present new EEP algorithm without self-adaptive strategy parameter. Experimental results show that this new EEP outperforms to other existing EP and obtains excellent high quality solutions with fine tuning parameter value.