A Multi Agent Scheme and Optimization for Big Data Management of Sensor Networks in Smart City Management

Because of complex sensor networks in smart city management, it is very difficult to optimize the data processing from all kinds of sensors. Here a multi-agent system (MAS) is made for data processing and optimization of sensor networks in smart city management. First, the sensor network in smart city management is modeled as a self-organized and decentralized agent swarm. In the MAS, each agent’s objective value is reckoned on-line and the best agent’s update rule is on the basis of proportional control concept. Second, each agent is organized by itself to herd to the prime agent in group. And when it avoids the crash between agent and the closest obstruction/agent, it moves to a moving target. Third, to analyze the MAS’s dynamics, the eigenvalue of time-varying discrete system’s analysis is made. Besides, a guideline is put forward for application on how to adjust the parameters of MAS’s. Finally, the results of the simulation verify that the proposed self-organized swarm system is effective in the capability of migration and flocking.