Presentation Title

Impacts of Negotiation in Central Place Foraging

Faculty Mentor

Jason Isaacs

Start Date

17-11-2018 8:30 AM

End Date

17-11-2018 10:30 AM

Location

HARBESON 16

Session

POSTER 1

Type of Presentation

Poster

Subject Area

engineering_computer_science

Abstract

Multi-agent systems (MAS) consist of a series of individually functioning computer systems that interact with each other in order to solve problems that may be more difficult or impossible for a single agent. One application of MAS is Swarm Robotics (SR), which is taking the basic principles of MAS and applying it to automated robots. One of the most prevalent sub-problems in SR is Central Place Foraging (CPF), which involves searching for, collecting, and the delivery of resources to a central home location. While it is logical to believe that the more agents there are contributing to the CPF task, the more efficiently the task can be completed; resource collection efficiency may in fact be hindered by the addition of more agents. The purpose of this research is to develop a mechanism for recruiting other agents to assist in the CPF task which allows the agents to negotiate with each other to distribute the work efficiently. Negotiation for our purposes is defined as an agent sending out a call for assistance when encountering multiple resources and all other agents negotiating amongst themselves which ones should assist, if any. The central question of this research is to test if such a recruiting mechanism which involves inter-agent negotiation can increase the CPF efficiency over more naïve recruiting mechanisms. To test this hypothesis, we will be running multiple simulations in which we will compare the collection efficiency of this negotiation algorithm as opposed to a more naïve algorithm in which the nearest free agent is recruited automatically.

Summary of research results to be presented

We will be presenting the results of several simulation trials which quantify the foraging efficiency of the proposed approach and compares to the results from baseline approaches

This document is currently not available here.

Share

COinS
 
Nov 17th, 8:30 AM Nov 17th, 10:30 AM

Impacts of Negotiation in Central Place Foraging

HARBESON 16

Multi-agent systems (MAS) consist of a series of individually functioning computer systems that interact with each other in order to solve problems that may be more difficult or impossible for a single agent. One application of MAS is Swarm Robotics (SR), which is taking the basic principles of MAS and applying it to automated robots. One of the most prevalent sub-problems in SR is Central Place Foraging (CPF), which involves searching for, collecting, and the delivery of resources to a central home location. While it is logical to believe that the more agents there are contributing to the CPF task, the more efficiently the task can be completed; resource collection efficiency may in fact be hindered by the addition of more agents. The purpose of this research is to develop a mechanism for recruiting other agents to assist in the CPF task which allows the agents to negotiate with each other to distribute the work efficiently. Negotiation for our purposes is defined as an agent sending out a call for assistance when encountering multiple resources and all other agents negotiating amongst themselves which ones should assist, if any. The central question of this research is to test if such a recruiting mechanism which involves inter-agent negotiation can increase the CPF efficiency over more naïve recruiting mechanisms. To test this hypothesis, we will be running multiple simulations in which we will compare the collection efficiency of this negotiation algorithm as opposed to a more naïve algorithm in which the nearest free agent is recruited automatically.