Method for quickly searching multiple shortest paths

Abstract

The invention discloses a method for quickly searching for multiple shortest paths, which belongs to the field of computer networks and solves the problem of calculating multiple (assumed to be K) shortest paths from a source node to each other node. In the prior art, classic methods such as Dijkstra and A* algorithms are used, which require multiple iterations and complicated calculations; at the same time, the calculated paths cannot effectively avoid loops, and the actual use value is low. The method proposed by the invention can obtain the K shortest paths from the source node to all other nodes at one time, and there is no loop. It is assumed that the water flow starts from the source node and flows at a constant speed, and starts to diverge after reaching a confluence point. Record the diversion information at the confluence point to obtain a shortest path from the source node to the confluence point. The confluence point records K pieces of diversion information and then stops the diversion; proceed in this way, when all confluence points record K pieces of diversion information, you can get The shortest K paths from the source node to all the sink points. The present invention can be used in logistics scheduling, GPS navigation and other practical applications.

Description

technical field [0001] The invention discloses a method for quickly finding multiple shortest paths, which can be used in logistics dispatching, GPS navigation, etc., and belongs to the field of computer networks. Background technique [0002] With the rapid development of e-commerce and the rapid expansion of logistics network, how to quickly deliver goods purchased by users and improve user experience has become the key to the further development of the logistics industry; among them, the path planning problem in express delivery is the core of increasing delivery speed and improving user experience question. The actual express delivery is often from one logistics collection center to multiple delivery points, and the shortest path to each delivery point needs to be calculated separately. At the same time, multiple shortest paths are also needed for selection in reality, thus resulting in single-source and multiple-destination K Shortest path problem. Of course, the sing...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to: when adopting classic methods such as Dijkstra, A* algorithm in the prior art to carry out many (assumed to be K) shortest path computations of source node to every other node, owing to need multiple iterations, cause The calculation is complicated; at the same time, the calculated path cannot effectively avoid loops, resulting in the problem of low actual use value

Images