Shortest Path Generation Research Articles

Effective route generation framework using quantum mechanism-based multi-directional and parallel ant colony optimization

Path planning in a complex network is a traditional and important research area in the optimization and machine learning fields. Several metaheuristics have been proposed; however, they are limited by the number of iterations, limited resources, and computation times. Therefore, in this study, the above issue is solved by integrating the existing metaheuristic algorithm with a prediction method. As a metaheuristic algorithm, ant colony optimization (ACO) is modified and applied. Instead of modeling ants pursuing higher pheromone densities, the proposed framework predicts pheromone traits using the intermediate pheromone density. As the pheromone volumes in the intermediate stage fluctuate, they are modeled using a quantum mechanism. Thereafter, the following pheromone traits are estimated using the modeled stochastic differential equation and Ito integral. The predicted pheromones are used for the parameters in the subsequent ACO iterations. To demonstrate the effectiveness of the proposed framework, the shortest path generation in several large-scale networks is provided. The proposed framework is considered a highly efficient metaheuristic framework with the integration of quantum mechanism-based prediction and existing metaheuristics.

An explicit shortest path generation scheme for point stabilization of the car-like robot in free space

This paper proposes a shortest path generation scheme for point stabilization of the car-like robot. An implicit shortest point stabilization algorithm has been suggested conceptually by Vendittelli. As an extension, we present an algorithm which is more explicit and faster than previous one. Our scheme provides exact equations for practical implementation. Moreover, its computation time is remarkably lower than the previous one for the case in which the final point changes frequently. We provide a proof of the optimality and validate the performance of our algorithm through the simulation and experiment. where v(t) ,ω (t) are a linear and an angular velocity. These two velocities are control inputs to the robot. We assume minimum turning radius as 1 and |v(t )=1 | which means that the robot can go both forward and backward with constant speed. Now, we say(x(t0) ,y (t0) ,θ (t0)) as an initial posture and (x(tf) ,y (tf) ,θ (tf)) as a final posture, where t0 and tf are an initial and a final time. We also denote (xf ,y f) as a final point.