Paper Title
Hybrid Swarm Intelligence Algorithm for Optimal Cognitive Architecture Design of Operators in Nuclear Power Plant

Abstract
A hybrid swarm intelligence algorithm is proposed to determine the near-optimal human cognitive performance of operators under various initiating events that would lead to accidents in Nuclear Power Plants. A three-parameter Weibull distribution model is used to estimate Human Error Probability (HEP), a measure of the Human Cognitive Reliability valuated under dominant cognitive processes associated with the task being performed by the operators during an emergency situation. In conventional Human Cognitive Reliability (HCR) model, the cognitive coefficients that depict the human characteristics are treated as constants. However the values of these coefficients that represent various operator behavior should be carefully assigned for better estimates of HEP. In this study, critical human cognitive factors such as attention, perception and memory under different cognitive levels are evaluated in order to design an optimal human cognitive architecture that minimizes the HEP. The results obtained from the proposed algorithm revealed that the hybrid swarm intelligence algorithm outperformed the generic Ant Colony Optimization (ACO) algorithm. Keywords - Cognitive Architecture Optimization, Human Error Probability, Hybrid Swarm, Intelligence Algorithm, Nuclear Power Plant.