Data Collection from Wireless Sensor Networks: OpenMP Application on the Solution of Traveling Salesman Problem with Parallel Genetic Algorithm and Ant Colony Algorithm

Abstract

Parallelization of algorithms can reduce time in many cases while using multiple cores at the same time. Although Algorithms such as Genetic Algorithm (GA) and Ant Colony (AC) are widely used optimization algorithms to solve the nonlinear problems it is usually time consuming. This study aims to solve a well-known NP-Hard problem, The Travelling Salesman Problem (TSP), by using both parallel and serial GA and AC. As an application, the data collected from the wireless sensors networks (WSNs) were used and the performance values of the running algorithms were compared. Reducing the travelling time in WSNs avoids losses in energy consumption caused by multi-tab transmission, but causes a long delay. Additionally, application was made with Open Multi-Processing (OpenMP) and its performance was compared with serial programming. According to the findings while both methods reduces the time in half when they run parallel, the performance of GA is much superior than AC.

Keywords

• genetic algorithm
• ant algorithms
• open mp
• travelling salesman