بهینه سازی شبکه جریان سریع بر اساس ضد بهینه سازی کلونی مورچه

There are fewer research results of optimizing the network fast flow by relative optimal path in the links. This paper presents a conception of link capacity based on the foundation of concept convergence expectation and convergence gradient. Based on the convergence gradient, the relatively optimal path is used to analyze problem in a reverse perspective of ant colony algorithm, which is also called anti-ant colony algorithm. According to the pheromone strength and rate of flow, the link makes a second judgment to decide the route. In this way, the load balance rate of the flow achieved and the congestion of the network can be avoided. Experimental results show that the method achieved link load balancing and the network resources utilization coefficient can be improved.

In the current development of practical engineering, users put forward their demand for a higher quality of service requirement. It becomes the primary target to reduce congestion between the points, efforts to improve the coefficient of resources utilization, and provide a higher QoS assurance for the terminal user[1]. These requirements requires not only broadband access networks but higher network bandwidth, have the ability to provide differential service to satisfy the demand of user and application [2]. From the current network technology development, the growth of network bandwidth is much higher than the growth in processor speed and memory access rate, server will become a bottleneck in the network transmission, via high-performance local area network or internet through the server cluster is a effective structure which can provide highly scalable and available network services [3]. The core network equipment already unable to afford the data flow and growing strength calculation of the current situation separately. So how to optimize the control of the flow reasonably is becoming a TE(Traffic Engineering) research hotspot [4-5].

In this paper, an anti-ant colony algorithm realize flow balancing based on the foundation by the introduction of convergence and convergence of expectations. Based on the pheromone on the link when ants find the optimal path to estimate the load condition on the network link, when the pheromone on the link is high allocate less flow, otherwise allocate much flow. The experiment result shows that via control the pheromone on the link can realize network load balancing, it can improve the utilization factor of the network in efficient, reduce network time delay, improve the utilization factor of the bandwidth, reduce packet loss. This paper solves network load balancing based on the foundation of convergence and convergence of expectations. This method take into account both the convergence rate and achieve network load balancing. The result of stimulate experiment shows that this method can judge the flow on the network link, then allocate the flow, achieve load balancing, improved the network efficiency.