Research On Artificial Bee Colony Algorithm

Swarm intelligence is one of the intelligent algorithms, which comes from the swarming behaviors of animal groups, has be used in solving optimization problems. It provides a new approach for solving optimization problems in many areas, such as management science, computer science, control engineering and other fields. Therefore, it has attracted more and more attention in recent years. Artificial bee colony algorithm is a novel Swarm intelligence algorithm, with simpler operation procedure, less control parameters and stronger robustness. Though it has been accepted and promoted by more and more scholars, when used in solving complex problems, it still encounters many problems, such as the low convergence rate and fall into local optimization. This thesis aims at improving the ABC algorithm by various strategies to enhance its ability for solving complex problems.First, aiming at improving the accuracy, local search capabilities and speed of operation, a quick self-adaptive artificial bee colony algorithm (QAABC) is proposed. This method improves the nectar update model of ABC algorithm, uses a new Scale Factor which can adaptive adjustment the search scope. Then, improves the way follower chose leader. Comparison of the performance of the proposed approach with Standard ABC and ABCP on5test functions is experimented, and later QAABC algorithm is applied to the engineering optimization design problems. The effectiveness of the improved algorithm is confirmed by the calculated results.Then, for the constrained problems, an improved algorithm named Guide And Follow Artificial Bee Colony (GFABC) is proposed. GFABC uses different methods for the two kinds of individuals (in the feasible area and out of the feasible area), and links them together. To be tested by eleven classical constrained problems, the GFABC shows to be a potential method to solve the constrained problems. Then the GFABC is applied to solve a real-world problem for the simplified alkylation process. The effectiveness and practicality of the improved algorithm is confirmed by the calculated results.