Pseudo Priority Queues for Real-Time Performance on Dynamic Programming Processes Applied to Path Planning

A family of approaches for the implementation of Dynamic Programming optimization can be improved for real time applications if the queuing stage is improved. The computational cost of maintaining a priority queue (PQ) for the standard Dijkstra’s algorithm has a theoretical lower bound that is function of the queue population. However a perfect PQ is not a necessary condition for implementing the Dijkstra’s algorithm. In many cases a weak version of a PQ, called in this paper Pseudo Priority Queue (PPQ) or previously known as Untidy Priority Queue, allows the Dijkstra algorithm to reach the optimal solution at a lower cost , i.e. having a complexity O(n). This paper discusses the performance of the PPQ and shows that it perfectly replaces a full PQ in order to allow a DP process to achieve the optimal solution at a lower computational cost. The approach is applied for motion planning for the real-time operation of a robotic platform.