Information Systems - Research Publications
Now showing items 1-5 of 5
The ShaMAN agent metamodel
In this paper we detail recent research on agent metamodels. In particular we introduce a new agent meta-model called ShaMAN created with a specific focus on computer game development using agent systems - an application domain that is fertile ground for agent-oriented concepts, methodologies and tools. ShaMAN was derived by applying the concept of Normalisation from Information Analysis against a superset of agent meta-model concepts from the meta-models investigated.
Coordination in Adaptive Organisations: Extending SharedPlans with Knowledge Cultivation
Agent-based simulation can be used to investigate behaviouralrequirements, capabilities and strategies that might be helpful in complex, dynamic and adaptive situations, and can be used in training scenarios. In this paper, we study the requirements of coordination in complex unfolding scenarios in which agents may come and go and wherethere is no fixed organisational structure, with an eye to developing asimulation framework that can be part of a training system in the domain of emergency management. We argue the need for an extensionto the SharedPlans formalism required to support the sharing of knowledge about a dynamically unfolding situation, specfically: who is in theteam, and who holds relevant knowledge. Our rationale for such an extension is presented based on a prior case study of a railway accidentand a further analysis of the coordination and communication activitiesamongst the disaster management team during its recovery. We concludethat in addition to the obligations imposed by the standard SharedPlansframework, agents in complex unfolding scenarios also need knowledgecultivation processes to reason about the dynamic organisational struc-ture and the changing world state. We briefly express the requirementsof knowledge cultivation as obligations that could be imposed on agents.
Augmenting BDI with relevance: supporting agent-based, pervasive applications
The potential of pervasive mobile systems lies in the integrationof applications and services available on mobile devices.These services must be built on the premises of context awareness,such that the application takes advantage of environmentalinformation to improve its performance. Thechallenge is to process more environmental information withthe limited computational resources available on mobile devices.Intelligent Agents have been proposed as a suitable architecturefor these tasks. Such agents are designed around adeliberation cycle. We show that one of the problems agentsencounter in highly dynamic environments is that this cyclemisses opportunities to react to the changing environment.For this reason we propose an extra module that augmentsagent architecures with a “context-observer”. Based on thestructures used in the deliberation cycle, this module evaluatesthe relevance of received data. It does this parallel tothe normal cycle. We illustrate how such an approach couldimprove the context sensitivity of the agent’s behaviour andthus the usefulness of the application.
Cybraries in paradise: new technologies and ethnographic repositories
(Mahwah, NJ: Lawrence Erlbaum, 2006)
Digital technologies have altered research practices surrounding creation and use of ethnographic field recordings, and the methodologies and paradigms of the disciplines centered around their interpretation. In this chapter we discuss some examples of our current research practices as fieldworkers documenting music and language in the Asia-Pacific region in active engagement with the cultural heritage communities, and as developers and curators of the digital repository PARADISEC (the Pacific and Regional Archive for Digital Sources in Endangered Cultures: <http://paradisec.org.au>). We suggest a number of benefits that the use of digital technologies can bring to the recording of material from small and endangered cultures, and to its re-use by communities and researchers.
Preliminary analysis of ANZDATA: experiment report
This report comprises a collection of experiments of applying data mining to ANZDATA. Section 2 presents the result of applying decision tree, emerging patterns (EP) and temporal ARs mining, and Section 3 reports the result of association rules mining on ANZDATA.