Collective decision-making by bee colonies as model for optimization - the OptBees algorithm

Abstract

This paper presents OptBees, a new bee-inspired algorithm for solving continuous optimization problems. Two key mechanisms for OptBees are introduced: 1) a local search step; and 2) a process of dynamic variation of the number of active bees that helps the algorithm to regulate the computational effort spent in the search and to achieve improved results. The performance of the algorithm was evaluated, in terms of global search, in all twenty-five minimization problems proposed for the Optimization Competition of Real Parameters of the CEC 2005 Special Session on Real-Parameter Optimization, held in the 2005 IEEE Congress on Evolutionary Computation (CEC). The results obtained show that OptBees is a fairly powerful search approach, presenting a superior performance to many of the competitors and other more recent algorithms also evaluated in these problems, including other bee-inspired approaches.

Bio-inspired multiobjective clustering optimization: A survey and a proposal

“…[71] cOptBees was first presented by Cruz et al [4] and the algorithm was applied to different clustering problems, being capable of finding optimal clusters, generating and maintaining the diversity of solutions, and finding the correct number of clusters. A parametric sensitivity analysis of cOptBees was presented by Cruz et al [72] and it was applied to training a Radial Basis Function (RBF) neural network.…”

“…[71] cOptBees-MO is summarized in Algorithm 1 and its main steps are detailed in the following sections.…”

“…The recruiters that were inhibited are classified as non-recruiters. [75] In cOptBees-MO the similarity degree between two bees is determined based on the objects classified in the same cluster: The greater the number of objects classified in the same cluster, the greater the similarity degree. If the similarity between two bees is greater than or equal to the inhibition radius ρ, which is an input parameter, the bee that explores the worst site is inhibited.…”

“…Activated bees are placed in a random position in the search space, i.e., the swarm size is updated dynamically in each iteration. [75] 6.4 Determining the scout and recruited bees The number of artificial bees attracted by each recruiter is proportional to the quality of the explored food sources. Thus, a percentage of bees that was classified as non-recruiters are classified as recruited and associated with each recruiter proportionally.…”

“…The number of scout bees is n s = n nr − n r . [75] 6.5 Recruiting The recruiters explore promising regions of the search space and perform a recruitment process to attract the closest bees to the sites they explore. The recruitment, implemented by equation (42) or equation (43) ⊗ is the element-wise product: [75] −…”