An iterative approach for makespan-minimized multi-agent path planning in discrete space

• Main Authors: Wang, Wenjie, Goh, Wooi Boon
• Other Authors: School of Computer Engineering
• Format: Article
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/101773; http://hdl.handle.net/10220/19812
• Institution: Nanyang Technological University

Makespan-minimized multi-agent path planning (MAPP) seeks to minimize the time taken by the slowest of n agents to reach its destination and this is essentially a minimax-constrained optimization problem. In this work, an iterative max-min improvement (IMMI) algorithm is proposed to approximate the optimal solution of the makespan-minimized MAPP problem. At each iteration, a linear maximization problem is solved using a simplex method followed by a computationally hard MAPP minimization problem that is solved using a local search approach. To keep the local search from being trapped in an unfeasible solution, a Guided Local Search technique is proposed. Comparative results with other MAPP algorithms suggest that the proposed IMMI algorithm strikes a good tradeoff between the ability to find feasible solutions that can be traversed quickly and the computational time incurred in determining these paths.