Distributed Task Replication for Vehicular Edge Computing: Performance Analysis and Learning-Based Algorithm

In a vehicular edge computing (VEC) system, vehicles can share their surplus computation resources to provide cloud services. The highly dynamic environment of the network makes it challenging guarantee task offloading delay. To this end, we introduce replication VEC where replicas are offloaded multiple at same time, and is completed upon first response among replicas. First, impact number on delay characterized, optimal approximated in closed-form. Based analytical result, design learning-based algorithm (LTRA) with combinatorial multi-armed bandit theory, which works distributed manner automatically adapt itself dynamics system. A realistic traffic scenario used evaluate performance proposed algorithm. Results show that, under our simulation settings, LTRA an optimized reduce average by over 30% compared benchmark without replication, time improve completion ratio from 97% 99.6%.