Colony Swarm Intelligence Network Visualization Test Platform Development

Swarm Intelligence (SWARM INTELLIGENCE) which was inspired by the bionics mechanism was proposed for the parallelism of distributed problem solving and complex optimization algorithm for a class of combinatorial problems. Since this distributed computing algorithm have the edges of the positive feedback, robustness and parallelism, etc, it has been widely used in the computer simulation, pattern recognition, data mining, network communications and many other fields. Ant colony algorithm is of the field of swarm intelligence. It has been proved to be an efficient optimization algorithm by more and more in-depth study. Ant colony algorithm is based on the foraging of ant colony which demonstrates the high level intelligent. It has established a mathematical model of mutual communication and coordination for the ant colony as a whole, and the model has been successfully applied to the traditional TSP problem to obtain the most satisfactory optimal solution.In this paper, the theory of swarm intelligence of ant colony algorithm and its basic principle of the mathematical model were discussed. Based on the study of the application of ant colony algorithm in the practical application, we point out some limitations of the application which exist by now and made some more effective improvements on the traditional ant colony algorithm theory refer to some other different intelligent algorithm. By the same time, In order to further expand the audience of ant colony algorithm, in this paper we also established a web-based ant colony algorithm test platform. This article mainly finished the following specific works:1. To resolve the problem that traditional mathematical models which ant colony algorithm relies on are too ideal, in this paper we take one kind of special situations where the individual in the ant colony is not completely controllable and the process of transportation may get into stagnant in any given time into consideration and present one dual-constrained ant colony algorithm to ensure the safety of the entire transportation system. This dual-constrained ant colony algorithm leads the concept of virtual break into the whole transportation network and modifies the distance between two different types of unreachable nodes in the network to reconstruct the connectivity of network. Under the premise of safety, we obtain the optimal solution which fully complies with the system requirements of such problems in this way.2. To resolve the problem that the ant colony algorithm is often prone to fall into local optimum, in this paper we introduce the idea of greedy algorithm into the applications of ant colony algorithm. According to the idea of local optimum in the greedy algorithm, this paper presents one kind of negative feedback ant colony algorithm based on minimum distance balance factor. The balance factor of the minimum distance optimal control strategy is used as a negative feedback which is attached to the ant colony algorithm for solving the problem and this hybrid algorithm improve the quality of optimal solution which the traditional ant colony algorithm may obtain.3. To resolve the problem that the strategy of local optimum in minimum distance balance factor algorithm may lead to one optimal solution of poor quality, this paper advanced the idea of estimates to judge the distance between the nodes. This idea expands the scale of information which minimum distance balance factor algorithm may access for computing in each step of the process and it can take more advantage of global optimization information, then based on this idea we present one improved algorithm for the minimum distance balance factor to improve the quality of the optimal solution and enhance the performance of process as a whole.4. This paper also established the INTERNET-based visualization of ant colony algorithm. The whole experiment environment was embedded in web pages and provided to end-users directly without downloading or installing. Users with any operating system can use any browser which supported JAVA plug-in to access the platform. The whole process of ant colony algorithm could be observed on this platform. The end users may carry some related qualitative research and quantitative analysis on this platform and it provides the possibility for the user to develop some new algorithm of their own on this platform.