Combinatorial optimization problems widely appear in various social and industrial applications, such as the vehicle routing and transportation problems. They are challenging to solve for traditional computing. Moreover, especially in edge devices, there is a high demand for in-time response and low power. However, the large energy consumption for these numerous operations and data fetches are the main obstacles for the hardware implementation of a problem solver. Our project will address the problem of time by using natural computing with parallel processing. Low-power techniques, such as approximate computing, are considered to boost computational and energy efficiency. This project provides a fresh perspective of how to improve the efficiency of solving combinatorial optimization problems. Taking the partner organization as collaborators, some findings will be submitted for publications or presentations. The intern is also willing to be included in the potential mentor group.