Martin J. Kollingbaum

Multi-agent coordination and control using stigmergy

This paper describes and discusses a novel design and a prototype implementation for manufacturing control systems, aimed at handling changes and disturbances. This novel design utilizes the concept of a multi-agent system. Agents in this system use an indirect coordination mechanism, called stigmergy. Stigmergy is a class of mechanisms that mediate animal-animal interactions. It consists of indirect communication that is taking place between individuals of an insect society by local modifications induced by these insects on their environment. The coordination mechanism in this paper is based on a technique used by food foraging ants. Food foraging ants provide the inspiration by the manner in which they spread information and make global information available locally; thus, an ant agent only needs to observe its local environment in order to account for nonlocal concerns in its decisions. A prototype was built to test the coordination technique. The prototype comprises a flexible manufacturing system model/emulation that has dynamic order arrival, probabilistic processing time, and some general perturbations such as machine breakdowns. The prototype served to investigate a specific research question: is it possible to create short-term forecasts based on the intentions of the agents. It has been intentionally kept simple to facilitate the understanding of what is happening in the system. Size and complexity of the prototype implementations are being augmented gradually in ongoing research.

Normative conflict resolution in multi-agent systems

Norms (permissions, obligations and prohibitions) offer a useful and powerful abstraction with which to capture social constraints in multi-agent systems. Norms should exclude disruptive or antisocial behaviour without prescribing the design of individual agents or restricting their autonomy. An important challenge, however, in the design and management of systems governed by norms is that norms may, at times, conflict with one another; e.g, an action may be simultaneously prohibited and obliged for a particular agent. In such circumstances, agents no longer have the option of complying with these norms; whatever they do or refrain from doing will lead to a social constraint being broken. In this paper, we present mechanisms for the detection and resolution of normative conflicts. These mechanisms, based on first-order unification and constraint solving techniques, are the building blocks of more sophisticated algorithms we present for the management of normative positions, that is, the adoption and removal of permissions, obligations and prohibitions in societies of agents. We capture both direct and indirect conflicts between norms, formalise a practical concept of authority, and model conflicts that may arise as a result of delegation. We are able to formally define classic ways for resolving conflicts such as lex superior and lex posterior.

Resolving conflict and inconsistency in norm-regulated virtual organizations

Norm-governed virtual organizations define, govern and facilitate coordinated resource sharing and problem solving in societies of agents. With an explicit account of norms, openness in virtual organizations can be achieved: new components, designed by various parties, can be seamlessly accommodated. We focus on virtual organizations realised as multiagent systems, in which human and software agents interact to achieve individual and global goals. However, any realistic account of norms should address their dynamic nature: norms will change as agents interact with each other and their environment. Due to the changing nature of norms or due to norms stemming from different virtual organizations, there will be situations when an action is simultaneously permitted and prohibited, that is, a conflict arises. Likewise, there will be situations when an action is both obliged and prohibited, that is, an inconsistency arises. We introduce an approach, based on first-order unification, to detect and resolve such conflicts and inconsistencies. In our proposed solution, we annotate a norm with the set of values their variables should not have in order to avoid a conflict or an inconsistency with another norm. Our approach neatly accommodates the domain-dependent interrelations among actions and the indirect conflicts/inconsistencies these may cause. More generally, we can capture a useful notion of inter-agent (and inter-role) delegation of actions and norms associated to them, and use it to address conflicts/inconsistencies caused by action delegation. We illustrate our approach with an e-Science example in which agents support Grid services.

Agent Support for Policy-Driven Collaborative Mission Planning

In this paper, we describe how agents can support collaborative planning within international coalitions, formed in an ad hoc fashion as a response to military and humanitarian crises. As these coalitions are formed rapidly and without much lead time or co-training, human planners may be required to observe a plethora of policies that direct their planning effort. In a series of experiments, we show how agents can support human planners, ease their cognitive burden by giving advice on the correct use of policies and catch possible violations. The experiments show that agents can effectively prevent policy violations with no significant extra cost.

Development of algorithms for agent based control of manufacturing flow shops

This paper reports on a development process for determining a suitable algorithm for application within agent based shop floor control systems. In particular, the algorithm discussed here is focussed on the achievement of high and consistent levels of production throughput in a manufacturing flow shop environment where there is a significant degree of flexibility associated with route selection. It is intended that in addition to performing well under nominal operating conditions, that such algorithms support a control system that is robust in the face of typical but unexpected process disturbances, such as congestion. It is emphasised that the algorithms are primarily intended for shop floor control and execution rather than scheduling for which numerous existing algorithms already exist.

Recovery mechanisms for semantic web services

Web service-based applications are widely used, which has inevitably led to the need for proper mechanisms for the web service paradigm that can provide sustainable and reliable execution flows. In this paper we revise recovery techniques in OWL-S and show how semantic annotations may ensure seamless web service provision in a sophisticated way, such as, exploiting the ontology-based description of processes in order to dynamically find alternative services as substitutes for failed services. We also discuss the consequences of these semantic-enabled approaches and point out required changes for integration in OWL-S.

Norm adoption in the NoA agent architecture

Agents based on reactive planning architectures use pre-specified plans as behaviour specifications. Normative agents are motivated by norms in their behaviour: obligations motivate them to act, prohibitions motivate them to refrain from certain actions, and permissions (or privileges) and capabilities specify the range of possible actions for such an agent. An important issue for normative agents is: under what circumstances is it appropriate for an agent to adopt a new set of obligations, prohibitions or permissions and what effect does it have on the agents normative state? In answering this question, a critical issue is whether or not this set of norms is consistent with the agents current normative state. Consistency of a set of norms is discussed in detail in respect to a reactive planning agent architecture -- NoA, but, it is argued, provides useful insight to the problem of norm consistency in general and particularly where practical reasoning agents are concerned.

Conflict resolution in norm-regulated environments via unification and constraints

We present a mechanism to detect and resolve conflicts in virtual environments, populated by agents whose behaviours are regulated by norms, that is, explicit representations of obligations, permissions and prohibitions. A conflict arises when an action is simultaneously prohibited and obliged/permitted. We use first-order unification and constraint satisfaction to detect and resolve such conflicts, introducing a concept of norm curtailment. We present an algorithm for norm adoption which preserves conflict-freedom. Our approach allows us to address indirect conflicts and conflicts arising from the delegation of actions between agents.

Norm Conflicts and Inconsistencies in Virtual Organisations

Organisation-oriented approaches to the formation of multi-agent systems use roles and norms to describe an agents social position within an artificial society or Virtual Organisation. Norms are descriptive information for a role --- they determine the obligations and social constraints for an agents actions. A legal instrument for establishing such norms are contracts signed by agents when they adopt one or more roles. A common problem in open Virtual Organisations is the occurrence of conflicts between norms --- agents may sign different contracts with conflicting norms or organisational changes may revoke permissions or enact dormant obligations. Agents that populate such Virtual Organisations can remain operational only if they are able to resolve such conflicts. In this paper, we discuss, how agents can identify these conflicts and resolve them.

Norm Adoption and Consistency in the NoA Agent Architecture

The behaviour of a norm-driven agent is governed by obligations, permissions and prohibitions. Agents joining a society or accepting a contract for the purpose of executing specific collaborative tasks usually have to adopt norms representing certain rules and regulations. Adoption of norms can cause problems – an agent maybe already hold norms that would be in conflict or inconsistent with new norms it adopts. How can it be shown that the set of norms is consistent to allow the agent to act according to the ideals that the norms specify? In general, the answer to such a question in a real-world situation is not simple. This paper addresses the problem of finding a pragmatic solution to the problem of norm consistency checking for practical reasoning agents in the context of the NoA Normative Agent Architecture.