Collaborative Agents - Research and Development: International Workshops, CARE@AI09 2009 / CARE@IAT10 2010, Melbourne Australia, December 1, 2009 and Toronto Canada, August 31, 2010, Revised Selected Papers
Abstract
This book constitutes the thoroughly refereed post-conference proceedings of the first two international workshops on computational models of collaboration in distributed systems: CARE 2009, held as satellite event of the 22nd Australasian Joint Conference on Artificial Intelligence AI09 in Melbourne, Australia, in December 2009 and CARE 2010, held in conjunction with the International Conference on Intelligent Agent Technology (IAT) in Toronto, Canada, in August 2010. The 12 revised full papers presented together with 2 invited lectures were carefully selected from a total of 45 submissions to both events. The workshops' thematic focus is on collaborative and autonomous agents that plan, negotiate, coordinate, and act under conditions of incomplete information, uncertainty, and bounded rationality.
Chapters (13)
In the multi-robot area coverage problem, a group of mobile robots have to cover an initially unknown environment using a sensor or coverage tool attached to each robot. Multi-robot area coverage is encountered in many applications of multi-robot systems including unmanned search and rescue, aerial reconnaissance, robotic demining, automatic lawn mowing, and inspection of engineering structures. We envisage that multi-robot coverage can be performed efficiently if robots are coordinated to form small teams while covering the environment. In this paper, we use a technique from coalitional game theory called a weighted voting game that allows each robot to dynamically identify other team members and form teams so that the efficiency of the area coverage operation is improved. We propose and evaluate a novel technique of computing the weights of a weighted voting game based on each robot’s coverage capability and finding the best minimal winning coalition(BMWC). Also we designed a greedy method and a heuristic method to find BMWC in O(n
log
n) time and O(n
2) time respectively. We tested these two algorithm with our base line method.
Warehouse management systems are traditionally highly optimized to a specific situation and do not provide the flexibility
required in contemporary business environments. Agents are advocated to provide adaptiveness and flexibility, and have been
used to solve specific problems in the warehouse logistics domain. However, for creating a general warehouse management system,
it is not clear for developers how and when to implement them. In this paper, we describe many of the design decisions when
constructing a warehouse management system, and illustrate how and when agents (and agent-organizations) are useful.
Crisis management is the process by which an organisation deals with a major unpredictable event that threatens to harm the
organisation or the whole society. Three elements are common to most definitions of crisis: (a) a threat to the organisation,
(b) the element of surprise, and (c) a short decision time [11]. Crisis management, or disaster management, is a relatively
new field of management. Typically, proactive crisis management activities include forecasting potential crises and planning
how to deal with them, for example, how to recover if your company’s computer system completely fails, but also how to evacuate
the city of Rotterdam in case of a major flood [10]. Hopefully, organisations have time and resources to complete a crisis
management plan before they experience a crisis. Crisis management in the face of a current, real crisis includes identifying
the real nature of a current crisis, intervening to minimise damage and recovering from the crisis.
The stable marriage problem has a wide variety of practical applications, including matching resident doctors to hospitals,
and students to schools. In the classical stable marriage problem, both men and women express a strict order over the members
of the other sex. Here we consider a more realistic case, where both men and women can express their preferences via partial
orders, i.e., by allowing ties and incomparability. This may be useful, for example, when preferences are elicited via compact
preference representations like soft constraint or CP-nets that produce partial orders, as well as when preferences are obtained
via multi-criteria reasoning. We study male optimality and uniqueness of stable marriages in this setting. Male optimality
gives priority to one gender over the other, while uniqueness means that the solution is optimal, since it is as good as possible
for all the participating agents. Uniqueness of solution is also a barrier against manipulation. We give an algorithm to find
stable marriages that are male optimal. Moreover, we give sufficient conditions on the preferences (that are also necessary
in some special case), that occur often in real-life scenarios, which guarantee the uniqueness of a stable marriage.
We present an agent-based model to compare different coordination patterns in joint fire support (JFS) scenarios. Modern warfighting
approaches depend heavily on a separation of concerns (like reconnaissance, coordination and engagement) and therefore impose
high requirements on the coordination of all involved parties. Following the General Reference Model for Agent-Based Modeling
and Simulation (GRAMS), we present an agent-based model of this problem domain. Our simulations indicate that decentralized
JFS coordination leads to smaller average times from identification of a target to final engagement, while at the same time
requiring extensive resources. Central coordination is more effective in terms of engaged units and reduced resource requirements,
but tends to take more time.
The Intelligent Collaborative Care Management (ICCM) project offers a comprehensive framework and architecture for understanding
and helping to manage the complete life cycle of customer care. This framework formally captures the main components of providing
a customer with a number of possibly interrelated services by various service providers. The customer has objectives expected
to be fulfilled by these services and the service providers themselves may have objectives in delivering the services to the
customer. These services are delivered over time and potentially the entire lifetime of the customer. The ICCM framework specifies
two functionalities: a) basic functionalities to generate care plans, form contractual relationships and deliver services,
including the specification of many complex interactions and constraints between service providers and customers, and b) extended
functionalities to support adherence of contract formation and service delivery and variations of contractual commitments.
We present first empirical results into the performance of a prototype system based on the ICCM framework. A key insight is
that mental models of human agents (i.e., service providers and customers) are required to efficiently intervene when contractual
commitment and service delivery processes go “off-track” and when the contract requires renewal and variation. Further, we
offer guidelines to optimise the benefit/cost trade-offs of applying interventions in achieving and managing relationships
and delivery.
This paper presents a geometric approach to multi-robots group formation with connectivity preservation (from a graph-theoretic
perspective) among group members. The controller demonstrates consistency among different formations, as well as stability
while performing dynamic switching between formations. Inter-robots collision avoidance is delivered through formation preservation,
while permitting high degree of formation re-adjustability. It has been proven that such formation approach would result into
complete, isomorphic formations (with regards to its first and second isogonic) with edge connectivity
l(G)=\frac14n(n-1)\lambda{(G)}=\frac{1}{4}n(n-1), and a unique, shortest connectivity link among group members. The complete connectivity along with the isomorphic property
of the formations would, in essence, not only guarantee that the communication among the robotic agents will be preserved,
but also relax the topological requirements for message passing among group members that might be needed while switching between
different formations. In addition, the existence of the inter-robot shortest connectivity link at the group level, would ease
the message routing once the information sharing among all the members of the group is necessary.
KeywordsCooperative Agents–Multi-Robot Systems–Group Formation
![Fig. 1. The interplay between norms, motivations and trust (adapted from [16])](https://www.researchgate.net/profile/Adel-Taweel/publication/227178087/figure/fig1/AS:393648601616395@1470864548624/The-interplay-between-norms-motivations-and-trust-adapted-from-16_Q320.jpg)
Just as in human societies, for which we have developed reasonably effective systems to organise and manage interactions in such a way as to minimise the impact of erroneous or malicious behaviour,
we also need to find ways to organise and manage computational entities in order to mitigate their potential deleterious effect
on computational systems. In this paper, therefore, we analyse the role of trust, organisations and norms in a motivation-based
view of agency that seeks to regulate behaviour, and illustrate some of these issues with aspects of several projects, including
the contract project, concerned with electronic contract-based e-business systems.
Simulations of consensus formation in networks of interacting agents have recently demonstrated that convergence to a small
number of opinion clusters is more likely when the network is adaptive rather than static. In order to better model realistic
social networks, we have extended an existing model of such a process, by the introduction of a parameter representing each
agent’s level of ‘authority,’ based on their opinion relative to the overall opinion distribution. Here we aim to determine
the importance of initial network topology for opinion formation in this model, using two distinct initial network topologies:
an Erdős-Rényi random network, and the Watts-Strogatz small-world network. It is shown that marked differences exist in statistics
of the model after opinion convergence. These include the number of interactions between agents needed to reach consensus,
as well as a clear influence of opinion tolerance on the network’s clustering coefficient, mean shortest path, and degree
distribution after convergence. This latter effect suggests some interesting possibilities regarding the topology of ‘converged’
networks.
KeywordsOpinion Formation–Consensus Formation–Adaptive Network–Social Network Analysis–Agent-based Model
The Minority Game (MG) is a simple model for the collective behavior of agents in an idealized situation where they have to
compete through adaptation for a finite resource. It has been regarded as an interesting complex dynamical disordered system
from a statistical mechanics point of view. In this paper we have investigated the problem of learning the agent behaviors
in the minority game. We assume the existence of one “intelligent agent” who can learn from other agent behaviors. We consider
two scenarios in this research: (1) Given an environment with complete information, i.e., all records of agents’ choices are
known to public. The intelligent agent can use a Decision Tree to learn the patterns of other agents and make predictions.
(2) If we only know the data of collective behaviors, we assume the data are generated from combining the behaviors of variant
groups of agents. The intelligent agent can use a Genetic Algorithm to optimize these group parameters in order to get the
best guess of the original system. The experimental results show that, in this configuration of MG in both environments with
complete information and incomplete information, the intelligent agent can learn from history data and predict which side
is the minority.
This paper discusses some of the recent cooperative multiagent systems work in the TEAMCORE lab at the University of Southern
California. Based in part on an invited talk at the CARE 2010 workshop, we highlight how and why execution-time reasoning
has been supplementing, or replacing, planning-time reasoning in such systems.
Effective decision making is vital in all healthcare activities. While this decision making is typically complex and unstructured,
it requires the decision maker to gather multi-spectral data and information in order to make an effective choice when faced
with numerous options. Unstructured decision making in dynamic and complex environments is challenging and in almost every
situation the decision maker is undoubtedly faced with information inferiority. The need for germane knowledge, pertinent
information and relevant data are critical and hence the value of harnessing knowledge and embracing the tools, techniques,
technologies and tactics of knowledge management are essential to ensuring efficiency and efficacy in the decision making
process. The systematic approach and application of knowledge management (KM) principles and tools can provide the necessary
foundation for improving the decision making processes in healthcare. The example of the orthopaedic operating room processes
will illustrate the application of the integrated model to support effective decision making in the clinical environment.
Keywordsintelligence continuum–healthcare operations–multi agent systems–knowledge management
We investigate three aspects of collaborative learning in uncertain environments - aversion to exploration, structure of the
collaborative network, and strategic policy. In uncertain environments, these three aspects are likely to operate simultaneously
whereas previous research at best has focussed on two, if not one, of these aspects. This paper introduces our research effort
to study phenomena of learning under these circumstances. We propose a model of the three aspects of learning and conduct
a simulation experiment using this model to analyse the effect of the three variables on learning performance and knowledge
development.
Keywordscollaborative learning–exploration and exploitation
The paper presents the results of a study on the possibility of taking into account the coordination of the work of a warehouse in supply chains with consumers and suppliers. At the same time, the main effective methods of logistic interaction are generalized and specified. The introduction describes the relevance and theoretical studies that serve as a prerequisite for the development of research into the problems of interaction between a warehouse and suppliers and consumers in supply chains. Further, the features of the logistical coordination of the work of warehouses in supply chains with consumers and suppliers are described. When modeling warehousing processes in supply chains, it is proposed to use indicators of the level of regulation of warehouse material flows when interacting with suppliers and consumers. These indicators also affect the value of the warehouse area, the method of determining which is also given in the paper. On the basis of the above methods, the modeling of logistics processes was carried out, and the optimal characteristics of warehouses in the supply chains in the interaction of suppliers and consumers were obtained. The resulting plans for warehouses and schemes for the movement of material assets were tested at an operating enterprise. At the same time, the simulation showed that a change in the level of logistic interaction between a warehouse and suppliers and consumers, reflected by the proposed indicators of the level of regulation of material flows of the warehouse, leads to a significant change in warehousing costs.
KeywordsWarehousePerformance coordination indicatorsSupply chainInteraction with suppliers and consumersModeling
For controlling the multi-robot formation with joining robots, an incremental centralized formation system is developed and a nonlinear model predictive control (NMPC) method is implemented as the controller. The incremental updating method is used to update the system's state in real time when there is new robot joining during formation process. Then a NMPC approach is developed to reformulating the formation system into a convex nonlinear minimization problem, which can further be transformed into a quadratic programming (QP) with constraints. Then, a general projection neural network (GPNN) is implemented for solving this QP problem online to get the optimal inputs. In the end, two examples of incremental multi-robot formation are performed to verify the effectiveness of this method. IEEE
Comprehensive models for organizations must, on the one hand, be able to specify global goals and requirements but, on the other hand, cannot assume that particular actors will always act according to the needs and expectations of the system design. Concepts as organizational rules (Zambonelli 2002), norms and institutions (Dignum and Dignum 2001; Esteva et al. 2002), and social structures (Parunak and Odell 2002) arise from the idea that the effective engineering of organizations needs high-level, actor-independent concepts and abstractions that explicitly define the organization in which agents live (Zambonelli 2002).
In the minority game (MG), an odd number of heterogeneous and adaptive agents choose between two alternatives and those who end up on the minority side win. When the information available to the agents to make their choice is the identity of the minority side for the past m days, it is well-known that the emergent coordination among the agents is maximum when [ ]. The optimal memory-length thus increases with the system size. In this work we show that, in MG when the information available to the agents to make their choice is the strength of the minority side for the past m days, the optimal memory length for the agents is always two (m = 2) for large enough system sizes. The system is inefficient for m = 1 and converges to random choice behaviour for [ ] for large N. Surprisingly, providing the agents with uniformly and randomly sampled m = 1 exogenous information results in an increase in coordination between them compared to the case of endogenous information with any value of m. This is in stark contrast to the conventional MG, where agent's coordination is invariant or gets worse with respect to such random exogenous information.
ResearchGate has not been able to resolve any references for this publication.