The Multi-Agent Programming Contest 2022: Coordinating Agents in a Dynamic World: Agents Follow the Rules, or Not (Lecture Notes in Computer Science, 13997)

✍ Scribed by Tobias Ahlbrecht (editor), Jürgen Dix (editor), Niklas Fiekas (editor), Tabajara Krausburg (editor)

• Publisher: Springer
• Tongue: English
• Leaves: 203
• Category: Library

✦ Synopsis

This book constitutes the 16th edition of the annual Multi-Agent Programming Contest, MAPC 2022.
It gives an overview of the competition, describes the current scenario. The first paper describes the contest in general and this edition in particular, focusing on the organizers’ observations. The following papers are written by the participants of the contest, describing their team of agents and its performance in more detail.

✦ Table of Contents

Preface
Organization
Contents
The Multi-Agent Programming Contest 2022
1 Introduction
1.1 Related Work and Competitions
1.2 Outline
2 The 2022 Scenario
2.1 Strictly Limited Number of Tasks
2.2 Addition of Roles
2.3 Applying Dynamic Norms
3 The Tournament
3.1 Participants
3.2 Warm-Up Phase
3.3 Results
4 Lessons Learned
5 Outlook
References
Optimization-Based Agents in the 16th Multi-agent Programming Contest
1 Introduction
2 An Overview of the System
3 Gathering Shared Knowledge
3.1 The Map of the World
3.2 Free Goal Cells and Goal Zones
3.3 Meeting Other Agents
4 Simple Path Finding
5 Managing Units
5.1 Assigning Agents to Units
5.2 The Flock of a Digger
5.3 Going to a Goal Zone
5.4 Computing the Constructor Position
5.5 Changing the Constructor Position
5.6 New Tasks
6 Collaboration Using Optimization Problems
6.1 Exploring the World
6.2 Working on Tasks
7 Analysing the Matches
7.1 Strengths and Weaknesses of the Approach
7.2 A Summary of the Matches
8 Possible Improvements
9 Conclusion
10 Team Overview: Short Answers
10.1 Participants and Their Background
10.2 Statistics
10.3 Technology and Techniques
10.4 Agent System Details
10.5 Scenario and Strategy
10.6 And the Moral of it is …
10.7 Looking into the Future
References
MMD: The Block Building Agent Team with Explainable Intentions
1 Introduction
2 Architecture
2.1 Agent Team Architecture
2.2 Software Architecture
3 Orientation
3.1 Perceptions and Observations
3.2 Map Building and the Dynamic Map
3.3 Agent Identifications
3.4 Map Merging
3.5 Looping Grid and Map Size Detection
3.6 Pathfinder
4 Team Coordination
4.1 Task Achievement
4.2 Role and Norm Management
5 Intentions
5.1 Common Intentions
5.2 Explorer Intentions
5.3 Task Achieving Intentions
5.4 Agent Intention Management
6 Debugging and Explanations
6.1 Challenges of the Debugging
6.2 Explanations
6.3 Logging
7 Match Analysis
8 Conclusion
A Team Overview: Short Answers
A.1 Participants and Their Background
A.2 Statistics
A.3 Technology and Techniques
A.4 Agent System Details
A.5 Scenario and Strategy
A.6 And the Moral of it is …
A.7 Looking into the Future
References
GOALdigger-AIG-Hagen Multi-agent System: Team Description
1 Introduction
2 Methodology
2.1 GOAL Language and Framework
2.2 Messaging Between Agents
2.3 Movement
2.4 Map Construction
2.5 Map Exploration
2.6 Agent Hierarchy and Roles
2.7 Task Selection with Machine Learning
2.8 Task Delivery
2.9 Sabotage of Rival Agents
2.10 Logging and Debugging
2.11 Tournament Preparation
3 Results
3.1 Effectiveness of the Saboteurs
3.2 Logging with the Mini-Percept
3.3 Tournament Matches
3.4 Effectiveness of the Ant Colony Optimisation
3.5 Effectiveness of the Machine Learning in Task Choosing
4 Conclusion
5 Future Improvements
References
General deSouches Commands Multi-agent Army for Performing in Agents Assemble III Scenario: FIT-BUT at MAPC 2022
1 Introduction
2 deSouches Multi-agent Architecture
2.1 Hierarchy of Goals and Missions
2.2 Top Level: deSouches/central Agent
2.3 Second Level: Organisational Agents
2.4 Third Level: Situated Agents
3 Multi-agent System Design for the Agents Assemble III
3.1 Strategy and Mission of the FIT-BUT Team for Agent Assemble III
3.2 deSouches in MAPC
3.3 Task Fulfilment by Agent Groups
3.4 Building Coalitions to Accomplish Tasks
3.5 Coordinator Agents and Missions in the System
3.6 Situated Agents Level
4 Evaluation of Participation in the Tournament
5 Conclusion
A Team Overview: Short Answers
A.1 Participants and Their Background
A.2 Statistics
A.3 Technology and Techniques
A.4 Agent System Details
A.5 Scenario and Strategy
A.6 And the Moral of it is …
A.7 Looking into the Future
References
The 16th Edition of the Multi-Agent Programming Contest - The GOAL-DTU Team
1 Introduction
2 The Strategy and Implementation of Our Agents
2.1 Discovery
2.2 Task Plans
2.3 Solving Tasks
3 Evaluation of Matches
3.1 GOAL-DTU VS GOALdigger
3.2 GOAL-DTU VS LI(A)RA
3.3 GOAL-DTU VS FIT But
3.4 GOAL-DTU VS MMD
4 Discussion
4.1 Strategy
4.2 Teams
5 Conclusion
A Team Overview: Short Answers
A.1 Participants and Their Background
A.2 Statistics
A.3 Technology and Techniques
A.4 Agent System Details
A.5 Scenario and Strategy
A.6 And the Moral of it is …
A.7 Looking into the Future
References
LI(A)RA Team - A Declarative and Distributed Implementation for the MAPC 2022
1 Introduction
2 The Agents Assemble III Scenario
3 Implementation
3.1 Technology
3.2 Methodology
3.3 Strategies
3.4 Tests
4 Results
5 Conclusion
A Team Overview: Short Answers
A.1 Participants and Their Background
A.2 Statistics
A.3 Technology and Techniques
A.4 Agent System Details
A.5 Scenario and Strategy
A.6 And the Moral of it is …
A.7 Looking into the Future
References
Author Index