A parallel two-stage genetic algorithm for route planning
H. David Mathias, Samantha S. Foley · 2020
Robotic path planning is an area of increasing importance given the interest in developing autonomous vehicles of all types and sizes. Consider, for example, a warehouse robot for a large internet commerce company. The robot's task is to retrieve a large number of items from warehouse shelves and deliver them to the packing area. Because efficiency is important, it is desirable for the robot to travel as short a distance as possible in completing its task. Though related to path planning, this is a distinct problem as there are multiple, perhaps many, destinations. We call this problem route planning. We present a parallel genetic algorithm that runs in two stages to solve the route planning problem. The two-stage approach significantly improves results over a similar singe-stage parallel genetic algorithm for this problem.