Simulation And Research On The Behavior Of Bee Swarm Based On Artifical Life

In this thesis, artificial life approach was used to the study of complex swarm behavior. Based on the analyzation and summarizing of various behavior of individual bees, we established bee's perceptual model, environment model, the mobile rules, foraging rules, obstacle avoidance rules, memory rules, communication rules and state transitions. Besides, a vivid, interactive, user-friendly 3D simulation platform was build using Java 3D, implementing the abstract bee colony rules. In the simulation we can find swarm intelligent behavior is emerged under the interaction of individual behavior. We focus on six representative experiments based on this simulation platform, including how the first bee to find food source, how to find better food source, follower bees adopt the greedy selection, a random selection and Roulette Wheel Selection method to choose a lead bee seperately, how the perception radius effect the simulation result, how the maximum detection time for scouts effect the simulation result, how the size of leaders and followers in the colony effect the simulation result. By repeatedly running the simulation program, we studied the complex behavior of bees by observing and analying the simulation results. Combining the simulation models, we specifically described the ideas and steps of the bee algorithm, and then applied it to design web crawler in vertical search engine. We introduced the idea of bee colony algorithm combining the global search strategy and local search strategy to vertical search engine web crawlers, making the web search more intelligent. Also detail technical details, including how to define seed pages, how to extract the contents of a page, how to calculate the degree relevant to the subject, how to collect keywords and organize them.