Improved Genetic Algorithm Based MEP Search Strategy for DSNs Intrusion Detection

In directional sensor networks (DSNs), the minimum exposure path (MEP) refers to the path with the lowest sensor monitoring intensity between any two target points. Finding MEP in DSNs is of great value to guide the coverage optimization and enhance intrusion detection capability of the network. Aiming at the key problem of how to search MEP more accurately in DSNs, we proposed a MEP search strategy based on improved genetic algorithm (IGA-MEP) while adopting the probabilistic sensing network model. Firstly, the mapping relationship between the uniform grid model and the weighted undirected graph model is used to limit the path trajectory of the target in the network, and the calculation formula of path exposure in the discrete domain is derived by combining the knowledge of graph theory. Secondly, in order to increase path diversity and ensure path continuity while using genetic algorithm solving MEP problem, adaptive crossover and mutation probability is introduced and its crossover and mutation method is also improved to propose an improved genetic algorithm (IGA). Finally, aiming at the path length optimization problem of MEP, we took consideration of the path smoothness into the objective function, and applied IGA to solve the problem, so as to propose IGA-MEP strategy. The simulation results show that the length of MEP found by IGA-MEP is reduced by 19.8% after the path smoothness function is introduced. In addition, compared with ant colony algorithm, original genetic algorithm and particle swarm algorithm, the minimum exposure value (MEV) found by IGA-MEP was reduced by 58.2%, 83.2% and 88.8%, respectively. Therefore, the IGA-MEP strategy proposed in this paper can solve the MEP search problem in DSNs with more accuracy.