Application of Artificial Capital Market in Task Allocation in Multi-robot Foraging

Abstract

Because of high speed, efficiency, robustness and flexibility of multi-agent systems, in recent years there has been an increasing interest in the art of these systems. Artificial market mechanisms are one of the well-known negotiation multi-agent protocols in multi-agent systems. In this paper artificial capital market as a new variant of market mechanism is introduced and employed in a multi-robot foraging problem. In this artificial capital market, the robots are going to benefit via investment on some assets, defined as doing foraging task. Each investment has a cost and an outcome. Limited initial capital of the investors constrains their investments. A negotiation protocol is proposed for decision making of the agents. Qualitative analysis reveals speed of convergence, near optimal solutions and robustness of the algorithm. Numerical analysis shows advantages of the proposed method over two previously developed heuristics in terms of four performance criteria.

Multi-Robot Item Delivery and Foraging: Two Sides of a Coin

“…In the absence of a centralized auctioneer, market-based approaches can be conducted in a decentralized manner [32]. Market-based approaches have been applied to multi-robot foraging and delivery problems [33][34][35][36].…”

Multi-UAV Cooperative Task Assignment Based on Orchard Picking Algorithm

“…The unmanned aerial vehicle (UAV) has the advantages of strong continuous combat capability, high maneuverability, and environmental adaptability due to the fact that they are not controlled by pilots [1][2][3][4]. In the process of actual combat, due to the complex and varied environment and wide distribution of targets, it usually requires multiple UAVs to cooperate with each other to carry out tasks [5], which makes the task execution time shorter and the completion rate higher [6,7]. Optimal decisions need to be taken for various mission-critical operations performed by UAVs [8].…”

“…Previous scholars have established various algorithms for the assignment of cooperative tasks to multi-UAVs. Author(s) [6][7][8][9], for example, proposed the fast task assignment (FTA) method algorithm based on Q-learning by neural network approximation and experience replay sequencing, which effectively transferred online computing to an offline learning process to assign tasks to heterogeneous UAVs under the condition of environmental uncertainty. A brain storm optimization was improved by local search procedure that each new candidate solution moves to the local best position * Corresponding author.…”