Research Of Novel Small-World Optimization Algorithm And Its Application For Complex Thermal Power Control System

Small-world network is one of forefront subjects and academic focus in the research area, has been widely used in many fields such as social networks, internet networks, biological products, transport networks and so on. Small-world network in optimization application is a new area and its research results are still rarely reported. In order to avoid the deficiency of current available small-world optimization algorithm, two novel real-coding small-world optimization algorithms were proposed and then applied to real-time optimal control of complex thermal system. The main contributions of this thesis are listed as follows:1. Small-world network theories and its important parameters were expatiated fully and systematically. By analyzing and summarizing currently available small-world optimization algorithms, generally existed problems in them were discussed in depth. These efforts were helpful for further study on the improvement and innovation of small-world optimization algorithm.2. Small-world optimization algorithm with binary-coding and decimal-coding were improved and a small-world optimization algorithm with random real-coding was proposed. Then the new algorithm was compared with the small-world optimization algorithm with binary-coding and the genetic algorithm in the simulation experiment of several international standard test functions. The simulation results indicate that the searching time can be significantly reduced by applying the new algorithm, furthermore the cumbersome and long-running problem of the currently available algorithm caused by encoding and decoding can be validly avoided. The simulation example that the new algorithm is successfully applied to real-time optimization of nonlinear system further verify its superior performance in industrial real-time control.3. The chaos theory was described briefly, and then the chaotic map was applied into the new real-coding small-world algorithm in order to generate the initial node set and make it have diversity. Meanwhile, chaotic search mechanism was introduced into the operator construction process by using the strong chaotic search ability in small space, thus a new chaotic small-world optimization algorithm was proposed. Function test and the example of control optimization for delay system show that the algorithm can effectively overcome the shortages of the currently available small-world optimization algorithm and greatly improve its optimization ability. 4. In view of time-varying characteristics and real-time stability requirements of supercritical unit, the small-world optimization algorithm with random real-coding was used for tuning parameters of RBF neural network and PID controller. Simulation results show that this strategy makes the control output of the main steam temperature can quickly track the input and access to good quality regulation under different loads.5. In view of time-varying parameters and higher real-time requirements of the response to load in thermal power unit coordinated control system, the chaotic small-world optimization algorithm was used for tuning parameters of DRNN neural network and PID controller to realize multivariable optimal control. Meanwhile, for the optimization problem with constraints in practical predictive control of the overheated steam temperature, the chaotic small-world optimization algorithm was used as the rolling optimization strategy to realize real-time predictive control. Simulation results show that the algorithm effectively meets the demands of real-time and fast stable for the control system.