Battery Aware Variable Voltage Task Scheduling for System Energy Consumption Minimization

Optimization of battery energy consumption by variable voltage task scheduling technique is studied in the paper. An efficient slack time allocation policy is firstly derived by comprehensive analysis on non-linear characteristics of battery. In further, two heuristic BATS (battery aware task scheduling) algorithms are proposed. The former takes the task discharge current as the priority function for slack time allocation, hence can reduce the peak power efficiently, while the later takes the task discharge current weighted by execution time as the priority function, hence can optimize the battery energy consumption more efficiently. Experiment results show that the proposed algorithms can greatly reduce the battery energy consumption by more than 30% and the peak power of the system simultaneously