Modelling the Dynamics of Intracellular Processes as an Organisation of Multiple Agents [pdf]
T. Bosse, C.M. Jonker, J. Treur
This paper explores how the dynamics of complex biological processes can be modeled as an organisation of multiple agents. This modelling perspective identifies organisational structure occurring in complex decentralised processes and handles complexity of the analysis of the dynamics by structuring these dynamics according to an organisational structure. More specifically, dynamic properties at different levels of aggregation in the organisational structure are identified, and related to each other according to the organisational structure. The applicability of this organisational modelling approach to address complexity in biological context is illustrated by a case study: the organisation of intracellular processes.
A Survey of Agent-Based Intelligent Decision Support Systems to Support Clinical Management and Research [pdf]
D. Foster, C. McGregor, S. El-Masri
The research into agent-based intelligent decision support systems is important to the medical industry because these systems can be used to improve the quality of healthcare in many ways. The areas these systems can be used in are diverse from the storage of medical records to the examination and evaluation of real-time data gathered from monitors. These systems are helpful to doctors and nurses in the diagnosis and treatment of patient with all kinds of conditions. This paper is a survey of the current research into agentbased intelligent decision support systems to support Clinical Management and Research, in order to first determine the current state of research in the area, second to help derive the key features and problems with these systems in the medical industry, and thirdly to use these key features and problems to explain how the current research would effect the development of a neonatal intensive care unit agent-based intelligent decision support system.
Managing Community Healthcare Information in a Multi-Agent System Environment [pdf]
R. Hill, S. Polovina, M.D. Beer
Most older, frail or disabled older people prefer to receive care in their own homes, in preference to the care provided within hospitals or residential home environments. Whilst the aims of community care address humanitarian issues, it is a challenging and expensive task to manage the organisation, logistics, quality assurance and efficiency of these services. UK National Health Service managers also recognise that hospitals can achieve a higher throughput rate of surgical cases if routine, rehabilitative care is delegated to the local community, removing it from the hospital ward. In recent years, UK Local Authorities have created policies that off load the delivery of care to private sector care providers. The community care scenario is thus an immensely complex and politically charged healthcare market place, freely trading care services in a competitive environment. In particular, contractual agreements between community healthcare agencies create a complex economic market place that must be described in a robust way if agent-managed services are to be accepted commercially. We consider the `Event Accounting' model in relation to the myriad of payment transactions within the community care environment, and propose a robust transaction-based framework for the management of community care Multi-Agent Systems, that attempts to address the gulf between abstract concept and low-level, multi-agent system implementation.
Utility-based Agents for Hospital Scheduling: A Conjoint-Analytical Approach [pdf]
H. Czap, M. Becker, M. Poppensieker, A. Stotz
The scheduling of centralized operating theatres in large hospitals can be regarded as an archetypal cooperative decision problem. Multi-agent Systems (MAS) form an appealing paradigm to support different variants of Computer Supported Cooperative Work, like the situation presented in this paper. In an MAS-setting, each involved individual can be represented by an intelligent software agent that carries the specific constraints and the main preference-structures of his principal. The scheduling can then be done by interagent negotiations, resulting in a cooperative solution, which optimizes "social welfare" and medical and organizational resource allocation simultaneously. One of the most serious problems encountered in using such a utility-based approach to multi-agent scheduling is the actual measurement of human preferences and utility functions. In this paper, we propose using conjoint analysis as a solution to this problem and introduce a software component that can be easily integrated into existing agent-applications.
Debian-Med Free Software in Health Care [pdf]
A. Tille
The development of Free Software has much in common with scientific research: the sharing of knowledge and to make progress.
Software in science co-evolves with data that is available to feed it. In the data driven molecular sciences, the information technology is particularly concerned to ease the data flow between applications. This is of particular importance because of the biological data's heterogeneity while individual entries are strongly semantically interdependent. Bioinformatics has developed technologies to communicate between data and tools.
With agent and grid technologies, Computer Science has developed means to operate across multiple databases, hereto also connecting otherwise independent institutes across the world. For the agent community for the access of standard technologies and databases, and for the grid technologies in particular, a common problem is the accessibility of information sources in respective local installations. These may differ in version, location or access permissions. Because of these differences, a common infrastructure still requires considerable local maintenance, particularly for the incorporation of novel data sources.
For computational grids, VOs determine installation paths and inter-institutional access permissions. The actual work for implementing such a agreements is imposed on the respective site\u2019s maintainer. A further unresolved issue is the heterogeneity of underlying hardware. This paper describes Debian-Med, a special interest group within the Debian Linux organisation, aiming to provide a hardwareindependent common view on Free software and databases for medical and biological research, and stresses its possible impact on the community as a backbone of grids and information agents in computational biology.
Immune System Modelling with Situated Cellular Agents [pdf]
S. Bandini, S. Manzoni, G. Vizzari
The Immune System (IS) constitutes the defence mechanism of higher level organisms to micro organismic threats. It is genuinely distributed system comprising mechanisms of adaptation to unknown threats by means of the cooperation of heterogenous autonomous entities, whose interaction endows the overall system with learning capabilities. This paper describes how a Multi Agent approach, and more precisely the Situated Cellular Agents (SCA) model, was applied to represent specific elements and mechanisms of the IS. After a brief description of the composing parts and internal mechanisms of the IS, the SCA model will be introduced and exploited to represent them.
Agent-Oriented Data Curation in Bioinformatics [pdf]
S. Miles
The practitioners of bioinformatics require increasing sophistication from their software tools to take into account the particular characteristics that make their domain complex. For example, there is a great variation of experience of researchers, from novices who would like guidance from experts in the best resources to use to experts that wish to take greater management control of the tools used in their experiments. Also, the range of available, and conflicting, data formats is growing and there is a desire to automate the many trivial manual stages of in-silico experiments. Agent-oriented software development is one approach to tackling the design of complex applications. In this paper, we argue that, in fact, agent-oriented development is a particularly well-suited approach to developing bioinformatics tools that take into account the wider domain characteristics. To illustrate this, we design a data curation tool, which manages the format of experimental data, extend it to better account for the extra requirements placed by the domain characteristics, and show how the characteristics lead to a system well suited to an agent-oriented view.
Simulation of mitochondrial metabolism using multi-agents system [pdf]
C. Lales, N. Parisey, J.P. Mazat, M. Beurton-Aimar
Metabolic pathways describe chains of enzymatic reactions. Their modelling is a key point to understand living systems. An enzymatic reaction is an interaction between one or several metabolites (substrates) and an enzyme (simple protein or enzymatic complex build of several subunits). In our Mitochondria in Silico Project, MitoScop, we study the metabolism of the mitochondria, an intra-cellular organelle. Many ordinary differential equation models are available in the literature. They well fit experimental results on flux values inside the metabolic pathways, but many parameters are diącult to transcribe with such models: localization of enzymes, rules about the reactions scheduler, etc Moreover, a model of a significant part of mitochondrial metabolism could become very complex and contain more than 50 equations. In this context, the multi-agents systems appear as an alternative to model the metabolic pathways. Firstly, we have looked after membrane design. The mitochondria is a particular case because the inner mitochondrial space, ie matricial space, is delimited by two membranes: the inner and the outer one. In addition to matricial enzymes, other enzymes are located inside the membranes or in the inter-membrane space. Analysis of mitochondrial metabolism must take into account this kind of architecture.
Gradually Intrusive Argumentative Agents for Diagnosis [pdf]
I.A. Letia
A new model of agent for acting in processes of diagnosis where more constraints are placed on the performance in real world domains is presented. Preferences on actions and policies are used to induce a preferred course of action by the diagnosis agent. Past experience, gathered from controlled trials, enables the experience agent to oer counterarguments for improving the diagnosis process. The outcome of actions are introduced in the scenario and agent models to allow argumentation on actions, not just beliefs. We argue that the approach is general and exible, allowing the natural integration of new knowledgeable agents, while, at the same time, oering human agents transparency of the decision process.
Engineering Agent-Mediated Integration of Bioinformatics Analysis Tools [pdf]
V.G. Koutkias, A. Malousi, N. Maglaveras
The availability of massive amounts of biological data, distributed in various data sources, has prompted the development of a wide range of data analysis tools. However, due to the heterogeneity and complexity of the domain, performing a specific analysis by combining tools is complicated. In this paper, we argue that the software agent metaphor constitutes a flexible approach towards integration of Bioinformatics tools and dynamic composition-execution of data analysis workflows. To substantiate this assertion, a scalable engineering methodology is presented, aiming to address simple as well as complex cases, in which generic Multiagent Systems are described to address different functional scenarios, incorporating agent coordination strategies and semantic domain descriptions. Based on this methodology, test cases are discussed on gene prediction tools to illustrate the applicability of the proposed systems.
Usinig a Personalized Adaptive and Cooperative MultiAgent System for Predicting Protein Secondary Structure [pdf]
A. Orro, M. Saba, E. Vargiu, G. Mancosu
In this paper, a generic architecture designed to support the implementation of applications agent-based, aimed at retrieving information among different Internet sources is presented. Information is filtered and organized according to personal interests explicitly stated by the user. User profiles are improved and refined throughout time by suitable adaptation techniques. The overall architecture has been called PACMAS, being a support for implementing Personalized, Adaptive, and Cooperative MultiAgent systems. PACMAS agents are autonomous and flexible, and could be personalized, adaptive and cooperative depending on the given application. The peculiarities of the architecture are highlighted by illustrating a relevant case study aimed at predicting protein secondary structures.
Validating MAS Models with Mutation [pdf]
E. Merelli, M. Young
A key reason for choosing agent-based simulation of a biological system over other kinds of simulation model is the potential for structural as well as behavioral correspondence between the simulation model and the modeled system. This correspondence both demands and makes possible new kinds of model validation.We propose evaluating model fidelity by introducing seeded faults that correspond to known or hypothesized mutations, thus tying software validation to biological mutation analysis. We illustrate with consideration of an agent-based model of carbohydrate oxidation in a cell.
Multi-Agent Simulation of Protein Folding [pdf]
L. Bortolussi, A. Dovier, F. Fogolari
A protein is identified by a finite sequence of amino acids, each of them chosen from a set of 20 elements. The Protein Structure Prediction Problem, fundamental for biological and pharmaceutical research, is the problem of predicting the 3D native conformation of a protein, when its sequence of amino acids is known. All current mathematical models of the problem are affected by intrinsic computational limits, and by a disagreement on which is the most reliable energy function to be used.
In this paper we present an agent-based framework for ab-initio simulations, composed by different levels of agents. Each amino acid of an input protein is viewed as an independent agent that communicates with the others. These agents are coordinated by strategic and cooperative higher level agents. The framework allows a modular representation of the problem and it is easily extensible for further refinements and for different energy functions. Simulations at this level of abstraction allow fast calculation, distributed on each agent. We provide an implementation using the Linda package of SICStus Prolog, to show the feasibility and the power of the method.