Routing problem in rectangular mesh network using shortest path based Greedy method

Abstract

A supercomputer can have thousands of processor-memory pairs which often referred as processing pins. Each of these pins is connected to each other through networks and passes message using a standard message passing mechanism such as Message Passing Interface. In this research, we consider the routing problem in rectangular mesh network. Each terminal pin in the network needs to be connected with its destination pin for it to function properly. Thus, maximizing the number of connection for each pair of pins and keeping the total energy throughout the network minimum becomes our main objective. In order to achieve this objective, each net need to be routed as shortest as possible. Therefore, developing a shortest path based routing algorithm is in need. In this research, Dijkstra’s algorithm is used to establish the shortest connection for each net. While this method guarantees to provide the shortest connection for each single net (if exists), however each routed net will become the obstacles and block later connections. This will add complexities to route later nets and make its routing longer than optimal or sometimes impossible to complete. Therefore, the routing sequence need to be rip-up and all nets need to be re-routed. This paper presents a complete routing algorithm which can further refine the solution by using Dijkstra’s based greedy method. The outcomes from this research is expected to benefit engineers from electric & electronic industry.

Obstacle Avoidance and Path Planning Methods for Autonomous Navigation of Mobile Robot

“…In the visibility graph, the obstacles are represented as polygons [ 50 ], and the vertical nodes of the polygonal obstacles are connected in such a way that the path length is minimized while the lines remain close to the obstacles. In contrast, the Voronoi graph uses the two closest points of the edges of the obstacles for planning and divides the domain into subdomains.…”

“…Then, based on Darwinian evolutionary theory, they generate a new population using the three genetic operators (mutation, crossover, and selection). The new populations are created until the stopping conditions are met [ 50 ]. Such stopping conditions are a time limit, the required fitness value, and the maximum number of generations.…”

A Review on Path Planning and Obstacle Avoidance Algorithms for Autonomous Mobile Robots

“…In the visibility graph, the obstacles are represented as polygons [15]. Te vertical of the polygonal obstacles have nodes connected.…”

“…An improvement of this method and its application in path planning have been presented in Fan and Shi [14], and it has been shown that the upgrades are reasonable and practical. Finally, in [15], the path planning problem in rectangular mesh networks was addressed by fnding the shortest link between each pair of sites using a Dijkstra-based greedy algorithm. [16] introduced the artifcial potential feld (APF) method for the mobile robotics problem by believing the divergences between attraction and repulsion forces help robot movement inside the environment.…”