Abstract:In multi-agent reinforcement learning systems, the actions of one agent can have a negative impact on the rewards of other agents. One way to combat this problem is to let agents trade their rewards amongst each other. Motivated by this, this work applies a trading approach to a simulated scheduling environment, where the agents are responsible for the assignment of incoming jobs to compute cores. In this environment, reinforcement learning agents learn to trade successfully. The agents can trade the usage right of computational cores to process high-priority, high-reward jobs faster than low-priority, low-reward jobs. However, due to combinatorial effects, the action and observation spaces of a simple reinforcement learning agent in this environment scale exponentially with key parameters of the problem size. However, the exponential scaling behavior can be transformed into a linear one if the agent is split into several independent sub-units. We further improve this distributed architecture using agent-internal parameter sharing. Moreover, it can be extended to set the exchange prices autonomously. We show that in our scheduling environment, the advantages of a distributed agent architecture clearly outweigh more aggregated approaches. We demonstrate that the distributed agent architecture becomes even more performant using agent-internal parameter sharing. Finally, we investigate how two different reward functions affect autonomous pricing and the corresponding scheduling.

Trading Agents for Roaming Users

Implementing an Agent Trade Server

Leveraging Workflow Interoperability Using Mobile Agent

Making Match Tradeoff System Based on Multi-agent in Mobile E-Commerce

Bargain Security Model Based on Mobile Agent

Multi-agent Interaction Technology for Peer-to-Peer Computing in Electronic Trading Environments

Agent trade servers in financial exchange systems

Decentralized scheduling through an adaptive, trading-based multi-agent system

Mobile Agent System For Supply Chain Management

Study on Agent-based architecture of reconfigurable SCM system

Design and Implementation of an Agent-Based Web Services Platform for Electronic Commerce

Consumer Oriented Supply Chain Management Based on Mobile Agent System

Policy-Driven Agent Social to Implement EBBP

Multiagent-Based Web Services Environment For E-Commerce

A Framework of Transferring Mobile Services with Agent Based Middleware.

MAgNET: mobile agents for networked electronic trading

Using Intelligent Agents to Understand Management Practices and Retail Productivity

Double Agent Architecture: A Framework for Personal Mobility

A Reflective Service Coordination Mechanism Based on Mobile Agents

Experiences with network-based user agents for mobile applications

Trading Strategy Based Portfolio Selection for Actionable Trading Agents.