T*: a weighted double-heuristic search algorithm to find the shortest path

Abstract

This paper proposes a weighted double-heuristic search algorithm to find the shortest path between two points. It can be used in numerous fields such as graph theory, game theory, and network. This algorithm, called T*, uses a weighted and heuristic function as f(x) = × t(x) + × h1(x) + γ × h2(x). It selects the path which minimises f(x) where x is a current node on the path, t(x) is cost of the path from start to x, h1(x) is a heuristic to estimate the cost from x to the straight line passing through start and target, and h2(x) is a heuristic to estimate cost of the cheapest path from x to target. Furthermore, , , and γ indicate effective weights of each sub-function on f(x). T* algorithm is compared to the Greedy and A* algorithms in terms of hit rate and the number of processed nodes. Comparison results show that the proposed algorithm has a high efficiency compared to other algorithms.

FUCBR: Fuzzy‐based unequal clustering and bat‐based routing protocol for WSNs assisted IoT

“…To solve diverse applications, numerous algorithms are proposed. [29][30][31] The bat algorithm 32,33 is among the classic swarm intelligence optimization techniques [34][35][36] that used the principle of echolocation of the bat to discover the optimum solution within the scope of the problem. Several researchers have proved this algorithm to verify its feasibility.…”

A dynamic bidirectional heuristic trust path search algorithm

“…In order to optimise the performance of A* algorithm, WA* [23] became the variant of A* by weighting the heuristic function h ( n ). Gharajeh [24] improved A* by introducing the weighted double‐heuristic function to find the shortest path and proposed the T* algorithm. It optimised the function of A* by adding three different parameters: to weight the cost of path from the source node to the current node, the estimate cost of the minimum cost path from the current node to goal node and the straight path through the start node and the goal node from the current node, respectively.…”

A many‐objective integrated evolutionary algorithm for feature selection in anomaly detection

“…21 The principle of feature selection is to improve the classification performance by simplifying the original dataset and select the optimal feature subset. Since the concept of swarm intelligence [22][23][24] was proposed by Hackwood and Beni in 1992, it has been widely used in various fields. At its core, several simple individuals form a group, which demonstrates advanced and complex functions through mechanisms such as cooperation, competition, interaction, and learning.…”