Research on Multi-Agent Collaborative Path Optimization Algorithm for Low Altitude Economic Logistics

Right now, the low-altitude economy is growing really fast, and figuring out how to plan delivery routes in this space has become more and more important. The usual path planning methods—like the A algorithm or genetic algorithm—don't work too well in this kind of fast-changing setting, especially when lots of drones are flying together. Problems like bad route choices, slow response times, and wasted resources happen a lot. That's why finding smarter and more flexible ways to improve drone logistics has turned into a hot research topic. In our study, we came up with a new method for optimizing drone paths that mixes teamwork between drones with reinforcement learning. We built a teamwork-style setup and used the deep Q network (DQN) to help drones work together, make better decisions, and change their routes when things around them shift. We tried this method out in a virtual test and saw clear improvements. It worked way better than the old methods. Things like route quality, time spent flying, energy usage, and overall system performance were all improved. One of the main things we added was the mix of teamwork between multiple drones and learning from experience, plus a special way to adjust flight paths on the fly in these low-altitude delivery cases. This gives a very hopeful solution for making multi-drone delivery systems smarter and more useful.