A Research On Parallel Ant Algorithm In Cluster Of PCs Environment

As the continuing development of computing technology, and emergence ofhighperformancecost efficient computers and highspeednetwork, the parallelcomputers are produced and put into practical use. The introduction of parallelprocessing into diversified researching field provides an efficient way to solve a seriesof problems.The parallel ant algorithm is an important method to solve the large TSP(Traveling Salesman Problem). And it is the lowcost,high efficiency projectproceeded in a cluster of PCs. This paper first introduced the basic theory of parallelalgorithms and then discussed the cluster of PCs system and MPI as well as the antalgorithm and TSP. According to this theory, an appropriate laboratory environment, asimple cluster of PCs based on Windows and MPI, was established. There are greatdifferences between the newly devised MPI Programming mode and the traditionalone. Secondly, this paper discussed the method how to proceed the MPI parallelprogramming in cluster environment and how to implement the MPI parallel antalgorithm. The time differences between Parallel ant algorithm and traditional onewere analyzed after this program was tested. In addition, the effects of parallelefficiency caused by network communication in parallel ant algorithm and traditionalone were also compared. Finally, from the theoretic and practical study, the feasibilityof proceeding parallel ant algorithm in cluster of PCs was discussed and meaningfulconclusions were also achieved.