Learning–interaction–diversification framework for swarm intelligence optimizers: a unified perspective

Due to the efficiency and efficacy in performance to tackle complex optimization problems, swarm intelligence (SI) optimizers, newly emerged as nature-inspired algorithms, have gained great interest from researchers over different fields. A large number of SI optimizers and their extensions have been developed, which drives the need to comprehensively review the characteristics of each algorithm. Hence, a generalized framework laid upon the fundamental principles from which SI optimizers are developed is crucial. This research takes a multidisciplinary view by exploring research motivations from biology, psychology, computing and engineering. A learning–interaction–diversification (LID) framework is proposed where learning is to understand the individual behavior, interaction is to describe the swarm behavior, and diversification is to control the population performance. With the LID framework, 22 state-of-the-art SI algorithms are characterized, and nine representative ones are selected to review in detail. To investigate the relationships between LID properties and algorithmic performance, LID-driven experiments using benchmark functions and real-world problems are conducted. Comparisons and discussions on learning behaviors, interaction relations and diversity control are given. Insights of the LID framework and challenges are also discussed for future research directions.
© 2018, The Natural Computing Applications Forum.