A Human Learning Optimization Algorithm with Link Prediction Strategy

This paper presents a novel human learning optimization algorithm with link prediction strategy (HLOLP), in which the neighborhood network topology update method for HLO is studied from the perspective of network topology dynamic evolution. Based on the link prediction principle, a link prediction index based on fitness is designed. Taking it as the basis of network topology updating, a new dynamic network topology updating strategy is constructed. According to the information in the learning process, the learning relationship between different individuals is dynamically adjusted, so as to improve the learning quality of individuals and the diversity of the population and enhance the ability of searching for the optimal algorithm. The presented HLOLP is applied to solve CEC14 benchmark function and its results are compared with other recent metaheuristic algorithms to evaluate its performance.