Study On The Distributed Autonomous Agent Optimization

In this thesis, the distributed autonomous agent optimization algorithm and its application are studied deeply based on artificial life and multi-agent theory. The main works are as follows:1. Modeling of distributed autonomous agent system.Based on the study of artificial life, evolutionary algorithm, multi-agent theory and autonomy oriented computing, in this thesis, we build a distributed autonomous agent (DAA) system framework by simulating the organization of human society and the reproducing, moving and dying behaviors of mankind. In the DAA framework, the attributes and behaviors of the autonomous agent, and the organization structure and communication mechanism among the autonomous agents have been defined, and the implementation steps of the proposed DAAS are also given.2. Function optimization based on DAA.Based on the proposed DAA framework, we propose a new distributed autonomous multi-agent optimization algorithm (DAO) and apply it to solve the optimization problem of complex function in the continuous space. Testing results by standard function show that the proposed DAO algorithm can find the optimal solution quickly and steadily.3. Face detection and segmentation based on DAA.In this thesis, we also study the skin-based face detection by the proposed DAA framework, and propose a new method to detect and segment human face in color image. Experimental results show that the proposed DAA-based method is superior to the conventional method which scans the whole image pixel by pixel in the aspects of robustness and efficiency.