Efficient Heuristic Algorithms for Single-Vehicle Task Planning With Precedence Constraints

This article investigates the task planning problem where one vehicle needs to visit a set of target locations while respecting the precedence constraints that specify the sequence orders to visit the targets. The objective is to minimize the vehicle's total travel distance to visit all the targets while satisfying all the precedence constraints. We show that the optimization problem is NP-hard, and consequently, to measure the proximity of a suboptimal solution from the optimal, a lower bound on the optimal solution is constructed based on the graph theory. Then, inspired by the existing topological sorting techniques, a new topological sorting strategy is proposed; in addition, facilitated by the sorting, we propose several heuristic algorithms to solve the task planning problem. The numerical experiments show that the designed algorithms can quickly lead to satisfying solutions and have better performance in comparison with popular genetic algorithms.

Automated summary

This article investigates a task planning problem, constructs a lower bound on the optimal solution based on graph theory, proposes a new topological sorting strategy, and designs several heuristic algorithms to solve the problem, which outperform popular genetic algorithms in numerical experiments.