Pursuit learning algorithm to minimize delay and energy consumption in vehicular-edge computing networks
Publication Name
Wireless Networks
Abstract
As vehicular technology improves every day, the number of computationally-intensive and power-hungry tasks that need to be computed by a vehicle has increased significantly. Vehicles cannot handle such tasks due to their limited computing resources and energy, hence it causes a challenge for vehicles to maintain their high performance. Vehicular-Edge Computing (VEC) evolution has been beneficial since vehicles can offload some of their tasks to the VEC servers, minimizing the vehicle’s energy consumption. However, it is crucial to ensure that vehicles can compute their tasks either locally within the vehicle or at one of the VEC servers to achieve the minimum total cost in terms of energy and delay. This problem is addressed by considering a system model for VEC networks, where vehicles can compute their tasks locally or at the VEC server attached to one of Road Side Units (RSUs). Then, the problem is formulated as a binary optimization problem, where the main aim is to minimize the total cost in terms of energy and delay. The optimization problem is then solved using the pursuit learning algorithm, which is a learning automata technique. This algorithm would give each vehicle the optimal offloading decision that would result in the minimum total cost for all vehicles in the network. The experimental results prove the proposed algorithm’s efficiency and fast convergence and show how it outperforms other existing algorithms by almost 15%.
Open Access Status
This publication is not available as open access