Research Of The Optimal Path Based On GIS

There’re lots of advantages to study the optimal path problem under the GIS circumstance, GIS can make the information visualization and process the spatial data efficiently, which make the results more efficieny and intuitive when studing the issues about optimal path problem. In this paper, the optimal path problem is about the logistics in the transport of dangerous goods, as it’s needs to consider safety factors about transport for dangous goods, the optimal path function can help vehicle drivers be familiar with path in advance, managers can monitor whether vehicle is traveling in accordance with established routes, which improve transportation safety and at the same time also save transportation costs. If vehicles were in dangour, the optimal path function can help acient to arrive sooner.The optimization route problem studied in this dissertation is solved by ACA (Ant Colony Algorithms). ACA features robustness, positive feedback and so on. It is easy to be implemented in programming, and integrated with other algorithms to solve various types of problems. But it has weak points like slow convergence speed, stagnation and falling into local optima. Improving the positive feedback effect of the algorithms can speed up the convergence speed, while it reduces the initiative of searching new route and make algorithms more prone to stagnant and fall into local optima.With the analysis of the pros and cons of basic ACA, this dissertation studies the pheromone updating rule and presents improved methods, which improve the convergence speed and little influence the initiative of searching new route. After the evaluation of the algorithms, the improved ACA can be proved better than before.The development is based on MapABC, Eclipse developing environment and Webwork-Spring-Hibernate framework. The API of MapABC provides enough functions, making the coding procedure easier. For instance it can effectively initialize the map, get the random coordinates on the map and calculate the routing distance rather than direct distance between two points. It also makes the calculation of routing distance between two random points in optimization route model more precise. Thus the data needed by algorithms model can be more accurate, and the technical supports for the optimization route coding can be more robust.