Resource pooling in vehicular fog computing

Chaogang Tang,Shixiong Xia,Wei Chen,Qing Li,Weidong Fang

doi:10.1186/s13677-021-00233-x

Abstract

Vehicular fog computing (VFC) provisions computing services at the edge of networks by fully exploiting the idle resources of vehicle loaded computer systems. Task scheduling and resource allocation revolved around VFC have gained tremendous attention recently. Currently, most of these works in VFC have focused on response time optimization or energy reduction. Computing services are provisioned in a pay-as-you-go model and vehicles as resource contributors are stimulated by the benefits obtained by leasing these resources. How to maximize their own benefits is one of big concerns but few of current works have recognized its importance in VFC. We in this paper introduce the notion of resource pooling into VFC where the computing resources of vehicles are pooled together to jointly provision computational services in a community. A genetic algorithm based strategy is proposed to solve the optimization problem for the sake of benefit maximization. Extensive experiments have been carried out to evaluate the approach and the numeric results have demonstrated that our strategy outstands other approaches with regards to the optimization objective.

Highlights

The Internet of Things (IoTs) defines a connection paradigm, where people and things are able to connect and communicate anytime, anyplace with anything and anyone, ideally using any network and any services [1, 2]
We in this paper introduce the notion of resource pooling into vehicular fog computing (VFC) where the computing resources of vehicles are pooled together to jointly provision computational services in a community [16]
genetic algorithm (GA) based decision making algorithm we present a GA based decision making algorithm (GADM) for the problem P and the corresponding pseudo code is shown in Algorithm 1

Summary

Introduction

The Internet of Things (IoTs) defines a connection paradigm, where people and things are able to connect and communicate anytime, anyplace with anything and anyone, ideally using any network and any services [1, 2]. Due to the wide deployment of RSUs, the number of communities which vehicles can join at the same time is explosively increasing It is not an easy work for vehicles to select the suitable community to contribute the computing services, especially considering the potential benefits. Comparison benchmark On one hand, as far as GADM itself is considered, several parameters can affect the performance of GADM to a great extend with regards to convergence rate, running time, optimal solution approximation and so on These parameters usually include the crossover probability, mutation probability, population size, and the number of iterations. Considering the mobility of vehicles in VFC, the response time can be acceptable

Conclusion

Findings

Availability of data and materials No