Energy Optimization on Joint Task Computation Using Genetic Algorithm

Joint computation is a form of collaborative job execution running at separate physical units, which are previously grouped by their unique functionalities. While existing studies have mainly utilized joint computation with direct coordination between nodes in different segments, it is worth considering another scenario where an additional node within a layer relays data to another layer. As a consequence, the node can serve as an aggregation point for data capture units prior to transmission to the sink node. However, this new arrangement produces additional transmission paths and can thus cause additional energy spending. This pilot study investigates the joint computation problem aiming at optimizing energy consumption. Relevant components, such as computation and communication, are taken into account and modeled into formal representation. A genetic algorithm-based solution is then used as a tool to optimize parameter setup. According to the experiment results, the metaheuristic algorithm has potential to achieve the optimal system configuration, emphasizing the data length that affects the final energy spending on communications. However, the algorithm cannot always guarantee the optimality as it relies on the random variable used in the process.