A greedy initialiser based meta-heuristic approach for planning unmanned aerial vehicles

Abstract The unpredictable benefits and low operational costs of unmanned aerial vehicles (UAVs) compared to their conventional counterparts caused a tremendous change in commercial and military concepts, and different solutions to problems regarding modern aerial vehicles were tried to further improve flight and job performances. One of the most challenging problems about the UAVs is known as the path planning problem, and a solution should satisfy some objectives related to the enemy anti-air weapons, fuel or battery consumption, and manoeuvre capability of the UAV being operated optimally. Immune plasma algorithm (IP algorithm or IPA) is a recent meta-heuristic optimiser, and its competitive performance has been validated over a set of engineering problems. In this study, a greedy initialiser that is responsible for generating a population of IPA was first introduced. Also, the treatment schema of the IPA was completely redesigned for more robust and detailed search characteristics without requiring either IPA-specific control parameters or their subtle configurations. The new IPA-based path planner, called greedy initialiser IPA (gintIPA), was tested by using three battlefields and 12 test cases belonging to them, and the obtained results were compared with the results of the well-known meta-heuristic techniques. Comparative studies showed that gintIPA is capable of planning more robust and safe UAV paths than other techniques for $91.6$ of all test cases. The proposed greedy initialiser gives a chance to start optimisation with qualified individuals and the newly designed treatment schema improving both exploitation and exploration routines significantly contributes to the gintIPA when outperforming other path planners.