Assessment of BAR: Breakdown Agent Replacement Algorithm for SCRAM

Assessment of BAR: Breakdown Agent Replacement Algorithm for SCRAM

Shivashish Jaishy (Aichi Institute of Technology, Toyota, Japan), Yoshiki Fukushige (Aichi Institute of Technology, Toyota, Japan), Nobuhiro Ito (Aichi Institute of Technology, Toyota, Japan), Kazunori Iwata (Aichi University, Toyota, Japan) and Yoshinobu Kawabe (Aichi Institute of Technology, Toyota, Japan)
Copyright: © 2017 |Pages: 17
DOI: 10.4018/IJSI.2017070101
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

In the Multi-Agent Systems, many agents work together towards achieving a defined goal. As it may be difficult for the agents to work in a dynamic environment, the current concept is trying to focus on the issues of situation where there may be cases of agent breaking down. This algorithm will distinguish and groupify the breakdown agents from the active agents. The authors are focusing on this scenario and replacement of breakdown agents by active agents by implementing the SCRAM- Scalable Collision-avoiding Role Assignment with Minimal-makespan, which has generalized to many Multi-Agent Systems specifically focusing on the collision avoidance among the agents. The authors are trying to address the impact and fate of breakdown agents, which otherwise is not yet addressed in SCRAM, through a new algorithm. This paper is designed to allow the generalization of the concept of SCRAM without any collision and disturbances even in the case of agent breakdown.
Article Preview

2. Role Assignment Problem

Scalable Collision-avoiding Role Assignment with Minimal-makespan for Formational Positioning (SCRAM) is a role assignment and positioning system that generalizes well to many realistic and real-world MAS (MacAlpine, Price, Stone, 2015). The key criterion of SCRAM is the distance that the agents travel until reaching their target positions. SCRAM scales up to thousands of agents with effective role assignment and positioning in such a way that there is a perfect mapping of agents with a collision free and minimal makespan (minimum time for all agents to reach their target locations). Minimizing the time for the agents to reach their target positions is a key factor in functioning when agents are moving to goal positions to complete a shared operation where all agents must be in place before the task can be completed and/or started. Such tasks include those requiring agents be coordinated when they start jobs at their target positions and environments.

Role assignment problem is the problem of task allocation to the different agents in MAS so that the agents from the current position could reach to the next position in minimum time also avoiding any collision. SCRAM has been shown to satisfy the following two properties so as to perfectly assign the roles to the agents, also known as being CM valid (Collision-avoiding with Minimal-makespan):

  • 1.

    Minimizing longest distance:

The longest distance from an agent to the goal, with respect to all possible mappings, is minimized. (Property 1)

  • 2.

    Avoiding collisions:

The agents move to their positions without colliding with any other agent. (Property 2)

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 5: 4 Issues (2017)
Volume 4: 4 Issues (2016)
Volume 3: 4 Issues (2015)
Volume 2: 4 Issues (2014)
Volume 1: 4 Issues (2013)
View Complete Journal Contents Listing