Abstract: This report presents an overview of the state-of-the-art methods and models for planning for teams of embodied agents. Due to the nature of the real world, this means we focus on multi-agent planning in stochastic, partially observable systems. In particular we focus on decentralized partially observable Markov decision processes (Dec-POMDPs), partially observable stochastic games (POSGs) and related models. Regarding such models, we review complexity results and recently proposed methods for finding (approximate) solutions.
Abstract: Future command teams will exists of human and artificial actors. This paper introduces
a taxonomy of collaboration types in human – agentteams. Using two classifying
dimensions, coordination type and collaboration type, eight different classes of human
– agent collaborations transpire. These classes might aid designers in pinpointing the
socio-technical design issues associated with these kinds of hybrid organizations.
Abstract: This paper describes the design, implementation, visualizations, results and lessons learned of a novel real-world socio-technical research system for the purpose of rescheduling train drivers in
the event of disruptions. The research system is structured according to the Actor-Agent paradigm: here agents assist in rescheduling tasks of train drivers. Coordination between agents is based on a team formation process in which possible rescheduling alternatives can be evaluated, based on constraints and preferences of involved human train drivers and dispatchers. The research system is the result of cooperation on decentralised multi-agent crew rescheduling between Netherlands Railways (NS) and the D-CIS Lab. The implementation is realized using the Cougaar framework and includes actual timetable and rolling stock schedule data and driver duty data.
Abstract: This paper describes the design of a socio-technical research system for the purpose of rescheduling train drivers in the event of disruptions. The research system is structured according to the Actor-Agent paradigm: Here agents assist in rescheduling tasks of train drivers. Coordination between agents is based on a team formation process in which possible rescheduling alternatives can be evaluated, based on constraints and preferences of involved human train drivers and dispatchers. The research aim is to explore
the effectiveness of a decentralized, flexible actor-agent based approach to crew rescheduling. The
research system is realized using the Cougaar framework and includes actual rolling stock schedule data and driver duty data. The current reduced-scale version shows promising results for the full-scale version end 2008.
Abstract: This research report discusses human group characteristics as a stepping stone to study human-agentteam characteristics and dynamics. A human-agentteam, or so called actor-agentteam (AAT) is a group of humans and agents who interact in a coherent and coordinated way towards a common goal. The concept of AATs relates to actor-agent communities (AACs), as AACs are groups of humans and artificial systems (socio-technical information systems) that intimately work together to achieve a common goal (i.e. solve a problem) (Iacob et al., 2009).
AATs are envisioned to increase human performance in (among others) safety and security domains, emergency management, and traffic control. However, the concept of AATs brings many challenges. Besides the realisation of agents as teammembers, and the realisation of real-world AATs, the interaction between agents and humans is a challenge. If agents are to become (task performing) group members, team membership requires much from agents regarding human-agent interaction. How should agents be designed to become teammembers in an AAT? How can humans best interact with agents? When do trust an agent, or rely on it?
This document discusses human group characteristics to draw implications for AAT dynamics. This document is a follow-up of Gouman et al. (2008) in which stages of team development, group membership and cohesion, subgroups, norms, roles, status, and leadership were discussed. The current report first addresses communication and decision making, after which team performance and implications for AATs are discussed.
Abstract: The collaboration between humans (actors) and artificial entities (agents) can be a potential performance boost.
Agents, as complementary artificial intelligent entities, can alleviate actors from certain activities, while enlarging
the collective effectiveness. This paper describes our approach for experimentation with actors, agents and their
interaction. This approach is based on a principled combination of existing empirical research methods and is
illustrated by a small experiment which assesses the performance of a specific actor-agentteam in comparison with
an actor-only team in an incident management context. The REsearch and Simulation toolKit (RESK) is
instrumental for controlled and repeatable experimentation. The indicative findings show that the approach is viable
and forms a basis for further data collection and comparative experiments. The approach supports applied actor-
agent research to show its (dis)advantages as compared to actor-only solutions.
Abstract: During a three month practical training period at the DECIS Lab the author investigated the use of multiagent planning frameworks for use with the RISK simulator. After evaluating a number of promising frameworks the Machinetta framework seemed the most promising. Machinetta was selected due to its extended use in various projects and its continued development. The author created a small agentteam that interacts with the simulated world in the RISK simulation environment and the actions of which are coordinated using Machinetta. With this agentteam a number of scenarios have been run to evaluate Machinetta. These scenarios show Machinetta to be a complex system offering many possibilities.
Abstract: This deliverable explores basic characteristics of human groups and teams in order to derive implications for actor-agentteams (AAT’s). From a socio-psychological group dynamics perspective group developmental stages, membership, cohesion, subgroups, social status, roles, norms and leadership are defined and explained in order to enhance the understanding of the processes that are part of human group behavior. The document subsequently briefly explains what ‘actor-agent-team’ means, making the assumption that the factors that play a role in human-only teams also play a role in AAT’s and putting further implications to discussion.
Abstract: In recent years, a number of research projects have been conducted to investigate how software agents can support collaboration among human crisis workers and assist them in performing their tasks. By allowing agents to assume the role of an equal team member instead of merely a supporting role, we could fully exploit the capabilities of agents. For example, wecould allow fire robots to take the place of firefighters for efficient fire ground operations in circumstances where fire fighters find difficulty in their activities because of
explosions, toxicleakage, heat, thick smoke hazards, etc. To allow agents in a team to be
treated as equal to humans, agents should behave a little like humans; issues associated with communication and coordination among team members become relevant as well.
In this thesis, we present a system for actor/agentteaming in simulated incident and crisis scenarios. In this system, agents are team members themselves, equal in status to human team members. The purpose of the system is to determine effectiveness of actor/agentteams compared to actor-only teams. Its main components are: a simulator called
RISK, Machinetta's proxy-based coordination framework which supports an actor/agentteam with coordination and planning facilities, and a communication framework that allows actors to communicate in natural language to their agentteam members. Actors and agents are connected to this system in order to collaborate in incident and crisis scenarios simulated by RISK. A pilot experiment has revealed important values that can be measured regarding team performance.
Abstract: The insertion of intelligent agents as active participants in an organization will
significantly alter team dynamics in general, and will transform standard command
teams into adaptive human – agentteams. In order to benefit from the adaptive
capabilities that this new type of organization promises, we need collaboration
schemes between human and artificial actors that harmonize with their respective
competences and demands. This paper explores various types of human – agentteams, with a focus on the issues
involved with designing dynamic task allocation and proper human – agent
collaboration. To this end, we introduce a taxonomy of human – agentteam types and
elaborate on the issues and challenges involved with each type.
Abstract: The DECIS Lab conducts research on Actor-Agent Communities, especially in the
domain of crisis response and management. The ICIS research program among others
hosts the SEAT (Sustained Effectiveness of Actor Agents Teams) research project in
which the researchers Dr. Niek Wijngaards, Dr. Masja Kempen and Dr. Manuela Viezzer
focus on the effectiveness over time of humans, agents, and teams composed thereof, in
difficult circumstances. Within this research project, theories and models are developed
which need to be tested in crisis situations. Decided was that ultimately a scenario
simulation toolkit is needed and that my project would explore and formulate
requirements for the use of scenarios and specifications of scenario aspects.
This report describes the results of this project. The scenario specification approach is
outlined and the (atomic) actions are described and illustrated in some detail.
Furthermore, this report describes the whole progress of the project. The student also
gives his personal evaluation about the project and the relevance to his education.
This graduation has several topics in which the student’s tasks and contributions lay.
These are each handled in detail further in this report.
• The student has created the game world concept which was based on the results
from the SES project. This resulted in a RDF specification. Details of these
specifications are found in chapter two.
• The student created several scenarios for a select number of researchers. These
scenarios have the intent to validate the researchers’ research and provide them
with some meaningful results. The student also created those scenarios in Game
, which functions as a simulation of the simulator.
• For future use, the student also created guidelines for scenario creation. These
topics can be found in chapter three.
• The student created a requirements document for the development phase of the
scenario toolkit. These requirements can be found in chapter four.
• Most importantly, the student also participated in the development process of the
RISK simulator toolkit, which resulted in consistency in design and construction
and that the original game world concept only received some limited changes.
This resulted in the development of the current version of the toolkit. This topic is
handled in chapter five of this document.
Abstract: Multi-agent systems are rapidly finding applications in a variety of domains, including robotics, distributed control, telecommunications, etc. Learning approaches to multi-agent control, many of them based on reinforcement learning (RL), are investigated in complex domains such as teams of mobile robots. However, the application of decentralized RL to low-level control tasks is not as intensively studied. In this paper, we investigate centralized and decentralized RL, emphasizing the challenges and potential advantages of the latter. These are then illustrated on an example: learning to control a two-link rigid manipulator. Some open issues and future research directions in decentralized RL are outlined.
Abstract: In actor-agentteams human and artificial entities interact and cooperate in order to enhance and augment their individual and joint cognitive ergonomic and problem solving capabilities. Also actor-agent communities can benefit from ‘ambient cognition’, a novel further reaching concept than ambient intelligence that hardly takes into account the resource limitations and capabilities changing over time of both humans and agents in collaborative settings. The Dutch Companion project aims at the
realization of an agent that takes advantage of the ambient cognition concerning actor-agent system dynamics such that natural emotion-sensitive interaction with an actor over a longer period of time can be sustained. We elaborate on our vision of
pursuing ambient cognition within actor-agent systems and present the plans and expected results of the Dutch Companion project.
Abstract: In actor-agentteams human and artiﬁcial entities interact and cooperate in order to enhance and augment their individual and joint cognitive ergonomic and problem solving capabilities. Also actor-agent communities can beneﬁt from ‘ambient cognition’, a novel further reaching concept than ambient intelligence that hardly takes into account the resource limitations and capabilities changing over time of both humans and agents in collaborative settings. The Dutch Companion project aims at the realization of an agent that takes advantage of the ambient cognition concerning actor-agent system dynamics such that natural social and emotion-sensitive interaction with an actor over a longer period of time can be sustained. We elaborate on our vision of pursuing ambient cognition within actor-agent systems and brieﬂy describe the goals of the Dutch Companion project.
Abstract: The main goal of this thesis is to provide a first overview of the current architectures the most able to design a cognitive agent. The notion of cognitive agent is in line with the Actor-Agent Community (AAC) project of D-CIS Lab (second part of this thesis). This project aims to design a prototype of an artificial system with cognitive capabilities (the cognitive agent) capable to interact with humans within a team (the Actor-Agent Community).
Abstract: Engineering complex highly-interactive systems consisting of both human and artiﬁcial agents (actor-agent communities) requires insight in the use and role of concepts by individuals and in interaction between individuals (both human and artiﬁcial). In this philosophically-oriented paper, the distinction between concept kinds is found to depend on processing differences for these kinds, rather than content-based or structural differences. In addition, this leads to the characterization of a new concept kind: affordance concepts. Our next step is to a) experiment with acquisition of affordance concepts and b) investigate the role of affordances for strategic management of teams.
Abstract: Thanks to advances in both computer science and
engineering, the divide between robotics and multi-agent systems
is shrinking. Robots are capable of performing an ever wider
range of tasks, and there is an increasing need for solutions
to high-level problems such as multi-agent coordination. In this
paper we examine the problem of finding a robust exploration
strategy for a team of mobile robots that takes into account
communication limitations.We propose four performance metrics
to evaluate and compare existing multi-robot exploration algorithms,
and present a role-based approach in which robots either
act as explorers or as relays. The result is a complete exploration
of the environment in which information is efficiently returned
to a central command centre, which is particularly applicable to
the domain of rescue robotics.
Abstract: This research report describes an experiment regarding assessing the effectiveness of actor-agentteams within the SEAT project in the CDM cluster in the ICIS research program. This document focuses on the background of actor-agentteaming, and on a methodology to assess the performance of an actor-agentteam in comparison with an actor-only team. The experimental design is described together with the measurements and analysis. The results show that the experimental setup using the REsearch and Simulation toolKit(RESK) provides a repeatable construct. The results of the current performance comparison show no large decrement; but also not a large increment in performance. This is mostly due to the current (low) level ofagent complexity, where improvements are needed in communication capabilities and (more) team-oriented helpful behavior.
Abstract: This white paper contains the report of a first try-out experiment regarding assessing the effectiveness of actor-agentteams within the SEAT project. This document describes the background, experimental design and measurements and analysis of a pilot experiment conducted in July 2008 with an actor-only team.
Abstract: The main objective of this document is to bring together our assumptions, models, definitions and other findings regarding actor-agentteams, and effectiveness thereof. As such, this whitepaper by the SEAT project contains our findings on sustained effectiveness of actor-agentteams of the first half of ICIS research programme. The foundations of our research reported herein are formed by three main topics: actor-agentteams, effectiveness, and methodology. This document contains material to gain insight in the width and scope of actor-agent communities, and should be treated as a starting point for additional research and explorations. These demarcations are necessary to restrict the amount of information and effort to manageable proportions.
Abstract: Urban Search and Rescue is a growing area of robotic research. The RoboCup Federation has recognized this, and has created the new Virtual Robots competition to complement its existing physical robot and agent competitions. In order to successfully compete in this competition, teams need to field multi-robot solutions that cooperatively explore and map an environment while searching for victims. This paper presents the results of the first annual RoboCup Rescue Virtual competition. It provides details on the metrics used to judge the contestants as well as summaries of the algorithms used by the top four teams. This allows readers to compare and contrast these effective approaches. Furthermore, the simulation engine itself is examined and real-world validation results on the engine and algorithms are offered.
Abstract: Crew rescheduling in response to disruptions is a difficult problem, due to the additional (social) constraints imposed on human workforce. In the real-world domain of train driver rescheduling in the Netherlands, an actor-agent based approach is taken to (a) support human dispatchers and (b) accommodate individual train drivers’ preferences. This paper outlines the task-exchange team-configuration process including the role of the
various rescheduling constraints. The rescheduling approach is designed for operation in a real world environment: to this end, a number of heuristics are discussed that are currently being examined to optimize the solution finding process with respect to
three dimensions: performance, quality and clarity. The heuristics have been implemented in a research system, supporting the full driver-agent population, working on real world data. This effort is an ongoing study on novel multi-agent approaches to crew
rescheduling, and is the result of cooperation between Netherlands Railways and D-CIS Lab.