RESEARCH ON NOISYS AND NOISY REDUCTION STRATEGY IN IEC ALGORITHM

The noises in the individual fitness evaluation will be unfavorable to the population evolution in IEC (interactive evolutionary computation), so it restricts extensive application of the algorithm in complicated optimization problem. This paper analyzes the sources of the noises in the IEC genetic algorithm, proposes a 3-phase noisy model based on the definition of the cognition evaluation and fatigue evaluation, then gives the cognition evaluation and fatigue evaluation description based on individual Hamming distance and noisy reduction strategy based on fitness confidence, and finally gives the evolving individual fitness adjustment algorithm.