Improved Ant Colony Algorithm And Its Application

Ant colony algorithm is a kind of optimization algorithm, which is made by the ants’ habits.The algorithm’s model is the process that the ants are searching for food source. The algorithm has many advantages, such as the simple structure, the strong robustness, implementation simplicity etc. The basic ant colony algorithm applies only on the discrete optimization, how to use the ant colony algorithm in the continuous optimization becomes a big problem.At first, the TSP problem is quoted to dwell on the ideal and the feature of the ant colony algorithm. Aiming at that the near-optimal solution may be existed and slow searching, it puts forward some improvement on the rules of ants’choice. The ants have two time-varying properties in the travel. Namely, how many cities the ants have been and how far the ants have gone. Here quotes this two properties to impact on the ants’choice, and then the improved result of the ant colony algorithm is verified by the simulation results.Secondly, this paper emphatically proposes an improved ant colony algorithm which used in the continuous domain. Based on the ant colony algorithm of the grid partitioning strategy, the improved algorithm is improved respectively in three aspects.1. After the grid meshing, the grids are uniformly distributed. Some point values of the continuous space are easy to miss. Because of the ergodicity and internal randomness of the chaotic phenomenon, the chaos sequence is introduced to make the distribution chaotic.2. The every spacing of the grids also have great influence on search precision, the every spacing of the grids are changed adaptively in this paper.3. Aimed at that it may have the optimal solution near the grids,"the local chaotic search" and "the superior selecting and the inferior eliminating" are quoted to avoid missing the optimal solution. The optimal problem of the function is a typical problem of the application and optimization. Four kinds of typical complex function are selected, through the simulation results, it is verified that the performance of improved ant colony algorithm is better.Finally, the researchers have always pay close attention on the practice of the continuous domains ant colony algorithm. The conventional PID control algorithm is not applied to the control system of the more complex model. The single neuron is integrated based on the conventional PID control algorithm, then the control performance of PID algorithm is improved.Fine alignment machine is a straightening device. While the load situation is different, the transfer function is different. So the corresponding transfer functions are made according to two situation of external load in the control system. Because of the complexity of the system, the general PID control cannot achieve certain requirements. The parameters of the single neural PID controller are set by the improved ant colony algorithm in the application of the control system of the Fine alignment machine. Through the MATLAB simulation result, the difference in control performance of the improved algorithm and the old one are compared.