A robotic Mobile Fulfillment System (RMFS) is a new type of parts-to-picker order fulfillment system where multiple robots coordinate to complete a large number of order picking tasks. The multi-robottaskallocation ...
详细信息
A robotic Mobile Fulfillment System (RMFS) is a new type of parts-to-picker order fulfillment system where multiple robots coordinate to complete a large number of order picking tasks. The multi-robottaskallocation (MRTA) problem in RMFS is complex and dynamic, and it cannot be well solved by traditional MRTA methods. This paper proposes a taskallocation method for multiple mobile robots based on multi-agent deep reinforcement learning, which not only has the advantage of reinforcement learning in dealing with dynamic environment but also can solve the taskallocation problem of large state space and high complexity utilizing deep learning. First, a multi-agent framework based on cooperative structure is proposed according to the characteristics of RMFS. Then, a multi agent taskallocation model is constructed based on Markov Decision Process. In order to avoid inconsistent information among agents and improve the convergence speed of traditional Deep Q Network (DQN), an improved DQN algorithm based on a shared utilitarian selection mechanism and priority empirical sample sampling is proposed to solve the taskallocation model. Simulation results show that the taskallocation algorithm based on deep reinforcement learning is more efficient than that based on a market mechanism, and the convergence speed of the improved DQN algorithm is much faster than that of the original DQN algorithm.
taskallocation enables heterogeneous agents to execute heterogeneous tasks in the domain of unmanned aerial vehicles, while responding to dynamic changes in the environment and available resources to complete complex...
详细信息
taskallocation enables heterogeneous agents to execute heterogeneous tasks in the domain of unmanned aerial vehicles, while responding to dynamic changes in the environment and available resources to complete complex, multi-objective missions, leading to swarm intelligence. We propose a bio-inspired approach using digital pheromones to perform scalable taskallocation when the number of agents, tasks, and the diameter of the communications graph increase. The resulting emergent behaviour also enables idle agents in the swarm to provide periodic or continuous connectivity between disconnected parts of the swarm. We validate our results through simulation and demonstrate the feasibility of our approach by applying it to the 3D coverage and patrol problem.
In this work we address the multi-robottaskallocation Problem (MRTA). We assume that the decision making environment is decentralized with as many decision makers (agents) as the robots in the system. To solve this ...
详细信息
ISBN:
(纸本)9783642130212
In this work we address the multi-robottaskallocation Problem (MRTA). We assume that the decision making environment is decentralized with as many decision makers (agents) as the robots in the system. To solve this problem, we developed a distributed version of the Hungarian Method for the assignment problem. The robots autonomously perform different substeps of the Hungarian algorithm on the base of the individual and the information received through the messages from the other robots in the system. It is assumed that each robot agent has an information regarding its distance from the targets in the environment. The inter-robot communication is performed over a connected dynamic communication network and the solution to the assignment problem is reached without any common coordinator or a shared memory of the system. The algorithm comes up with a global optimum solution in O(n3) cumulative time (O(n2) for each robot), with O(n3) number of messages exchanged among the n robots.
暂无评论