Simulating Agent Societies with PRESAGE Brendan Neville and Jeremy Pitt Intelligent Systems & Networks Group, Dept. of Electrical and Electronic Engineering Imperial College London, SW7 2BT, UK [email protected], [email protected]

Abstract. PRESAGE is a simulation platform for rapid prototyping of Agent Societies. It enables designers to investigate the effect of agent design, network properties and the physical environment on individual agent behaviour and long-term collective global performance. This paper reviews a series of experiments in various domains, including e-commerce, ad hoc networks, and colored trails.

1

Introduction

PRESAGE is a simulation platform for rapid prototyping societies of agents. It offers designers a flexible and generic set of Java classes, interfaces and tools with which they can simulate agent societies. This facilitates systematic experimentation over the parameter space to determine the effect of agent design, network properties and the physical environment on individual agent behaviours, social relationships between agents, and long-term collective global performance. We have used it to study the socio-cognition of components, the evolution of network structures, and the adaptation of conventional rules, in various domains, including e-commerce, ad hoc networks, and intellectual property rights. It is relatively straightforward to develop a prototype in PRESAGE. It is necessary first to define agent participant types: this can be done by extending the abstract class supplied with PRESAGE (to guarantee compatibility with the simulation calls and provide core functionality like message handling etc.) or by defining a new class. Then the network properties and physical world are defined, using or extending the given base classes. Finally, additional plugins and event scripts can be written for a number uses including altering paramenters, visualisation, data logging and user interface. The PRESAGE architecture is illustrated as a software stack (Fig. 1) depicting the base simulation module, the interfaces and abstract classes, simulation managers, and the platforms connectivity to external processes. Above this we have given some examples of how the user could utilise each of the classes and modules e.g. an auction scenario operating over a unstructured P2P network with/without a physical world. As PRESAGE is a platform for developing testbeds, it is neutral to the specific scenario, enabling the platform to be used in a variety of agent based experimental tasks. In this “demo paper”, in the next section, we illustrate several applications of the platform in independent testbeds for experimenting with agent societies.

Interfaces & Abstract Classes

PRESAGE

Managers

Initialisation

 Experimental Data Storage  Participant Data Structures  Relational Queries

Event Script Manager

 External Agents  Additional Simulators  RMI

Physical World

Example Connectivity

 Log Data  Real-time Graphs  Monitor Network Structure

Network

 Activate/Deactivate Participants  Change Participant Attributes  Cause Physical World Events

 Unstructured P2P  MANET  Hybrid P2P

Participant

Example Scriptable Methods and Plugins

 Tile World  Brownian Motion World  No Physical World

 Buyer  Seller  Auctioneer

Example Instantiations of Base Classes

Plugin Manager

TCP/IP Comms

MySQL Connection

External Connections

Simulation Loop

f

Core

unctions

Fig. 1. Representation of the platform Architecture

2

Demonstrator Applications

PRESAGE has been used in the following experiments with agent societies: Trust and Economics in Decentralised Trading: In this application of PRESAGE we study the use of socio-cognitive theories of trust and reputation for regulating behaviour in an agent-mediated marketplace which was both unmoderated and dynamic [?]. As well as demonstrating how distributed socio-cognitive mechanisms could approximate centralised solutions, this work illustrates the platform’s ability to support complex interactions between heterogeneous agents in the context of a simulated e-commerce environment. Coloured Trails and IPR: To demonstrate PRESAGE’s physical world environment we have developed an agent society version of the Harvard Coloured Trails [?] scenario, with colours and trails serving as an abstract representation of Intellectual Property (IP) concepts. We are using this test environment to investigate the provision of legal services such as Regulatory Compliance, Alternative Dispute Resolution and Law Making in agent societies (see Fig. 2a). Collaborative information filtering on unstructured P2P networks: PRESAGE is used to simulate an agent society whereby peers self-organise the P2P overlay network based on their peer models such that the network itself becomes a collaborative information filter. This has been used to investigate the effect of different agent preference distributions and peer selection models on the structure of the emergent network and the population’s welfare distributions. Resource management: In certain types of network, for example in virtual organization and ad hoc networks, it is a common problem to define a ‘fair’ collective decision-making mechanism for the allocation of a common resources and to

safeguard against either the over exploitation or under utilisation of resources. PRESAGE has been used as the basis of three investigations in this domain: – Local inter-agent negotiation, combined with sanction and trust mechanisms, have been used to provide pareto optimal resource utilisation, diminishing the effects of illegal and anti-social behaviour on the collectives welfare. – A simulation was developed to investigate the use of a voting protocol to address the problems of hand-over and cluster formation in MANETs. This demonstrated how service delivery can be maintained in a MANET, where only a fraction of the society may be present, there is no centralised record of the decisions, and all the network nodes change over time (see Fig. 2b). – PRESAGE has been to investigate the adaptation of rules over time to optimise resource management in a network environment which is highly volatile.

(a)

(b)

Fig. 2. a) Colored Trails and IPR tile world b) Voting in MANET’s

3

Summary

We have described the agent society simulation platform PRESAGE, and application domains which can be demonstrated. More information is in [?]. This work has been undertaken as part of the EU-funded ALIS project(IST 027968).

References 1. Neville, B., Pitt, J.: A simulation study of social agents in agent mediated ecommerce. In: Proceedings of the Seventh International Workshop on Trust in Agent Societies. (2004) 2. Gal, Y., Grosz, B.J., Kraus, S., Pfeffer, A., Shieber, S.: Colored trails: a formalism for investigating decision-making in strategic environments. In: IJCAI Workshop on Reasoning, Representation, and Learning in Computer Games. (2005) 3. Neville, B., Pitt, J.: Presage: A programming environment for the simulation of agent societies. In: Proceedings AAMAS Workshop on Programming Multi-agent Systems (ProMAS), Springer-Verlag (2008)

Simulating Agent Societies with PRESAGE

Abstract. PRESAGE is a simulation platform for rapid prototyping of. Agent Societies. It enables designers to investigate the effect of agent design, network properties and the physical environment on individual agent behaviour and long-term collective global performance. This paper reviews a series of experiments in ...

899KB Sizes 2 Downloads 160 Views

Recommend Documents

Simulating Computational Societies
and implementors, this can be used to model, analyse and explain the .... model and the software developed to facilitate experimentation with agent so-.

Simulating Reflector Antenna Performance with GRASP9 - GitHub
Surfaces. – conic sections (e.g. parabolas, ellipsoids, etc.) – point cloud (e.g. shaped surface). – planes. – struts. – surface with errors. • Rim defined separately.

Monitoring with Zabbix agent - EPDF.TIPS
server cache, and two minutes until agent would refresh its own item list. That's better ...... You can look at man snmpcmd for other supported output formatting ...

Simulating Focal Plane Array Observations with MeqTrees - GitHub
AzEl observation tracking a fixed offset position. • What's Next? ... Implement model in software using tree structure ... El = elevation of field or tracking centre.

Multi-Agent Search with Deadline
Jun 15, 2011 - continuous probability density function f whose support is X. .... The shaded area shows the acceptance set A(t), whose barycenter with.

Multi-Agent Search with Deadline
Jan 16, 2015 - We study a multi-agent search problem with a deadline: for instance, the situa- ... a predetermined deadline by which a decision has to be made. ..... bargaining solution irrespective of the distribution of offers, as the ... 9The “F

Simulating the Human Brain - Cordis
Understand the brain at all levels of organization (genes to whole brain); simulate the brain ... Build software applications to model, simulate, visualize and diagnose biologically ... ICT methods for pharmaceutical companies. (disease and drug ...

Animating a Conversational Agent with User Expressivity
Hartmann et al. [1] that are based on the wrist movement in 3D space, irrespective of joint angles (shoulder, elbow, etc.) information. In this work, we estimate spatial extent (SPC) and temporal extent (TMP) of the Hartmann's et al. [1] expressivity

Simulating the Human Brain - Cordis
Build a suite of analytics applications to process brain data. (signal analytics, visual analytics, real-time analytics, auto- analytics); build data display applications ...

Simulating Adaptive Communication
N00014-95-10223 to John Anderson at Carnegie Mellon University. ..... In fact, even when given explicit instructions that the computer could not ...... Computer Science and Engineering, Oregon Graduate Institute of Science & Technology.

Simulating Time in jsUnit Tests
Oct 2, 2008 - Sometimes you need to test client-side JavaScript code that uses setTimeout() to do some work in the future. jsUnit contains the Clock.tick() method, which simulates time passing without causing the test to sleep. For example, this func

registered societies Accounts
Apr 12, 2016 - may not always be in the best interest of the District and/or its schools to accept such funds. In ... April 2016 .... Social Host Liability Guideline.

Short Selling China - Societies
Feb 12, 2014 - China Inc. is changing. – All about “who” you know. – Foreign .... Chinese 'Apple' and 'IKEA' stores. – Shopping strips with Starbocks, King ...

Simulating Reach Motions
technique to fit polynomial equations to the angular displacements of each ... A description of current research on human motion simulations at the University of.

Living with Kin in Lowland Horticultural Societies - Semantic Scholar
Jan 24, 2013 - We use information technology and tools to increase productivity and facilitate ..... kin counts and Helm's measure, match to some degree with.

Simulating the Ionosphere - GitHub
Sep 30, 2009 - DEFINITION: Approximating measurements at intermediate scales/positions from scattered measurements. We have sparse measurements.

human societies
of forces that translate into the observable behavior of com- mon substances. .... A very heterogeneous fluid may mix under one of two conditions: 1) by ...

Short Selling China - CFA societies
Feb 12, 2014 - All about “who” you know. – Foreign ... Recognized social financing comprises 194% of 2013 GDP .... Source: China Media Project. Sources: 1) ...

PRESAGE: A Programming Environment for the ...
simulation model and core modules are discussed in section 3. In section 4 ... imation environment for agent societies we have paid particular attention to developing ... It can also call external applications for example. Gnuplot to .... unstructure

Simulating Frontotemporal Pathways Involved in ... - eScholarship
experimental manipulations. The current ..... Using this simulated manipulation of frequency and context, we .... IEEE International Joint Conference on Neural ...

Simulating Stochastic Differential Equations and ...
May 9, 2006 - This report serves as an introduction to the related topics of simulating diffusions and option pricing. Specifically, it considers diffusions that can be specified by stochastic diferential equations by dXt = a(Xt, t)dt + σ(Xt, t)dWt,

Gaming and Simulating EthnoPolitical Conflicts
is an editable list of norms/value systems from which each group's identity is drawn. The range across .... and cultural norms (Hermann, 1999), plus the additions of protocol vs. substance, and top level ..... Lsim Legend: Blue = Leader and Cops/Arme

PRESAGE: A Programming Environment for the ...
the environment serves as a rapid prototyping tool for agent societies, ..... cut methods to perform queries and updates, in addition to managing the java-sql.

simulating anisotropic frictional response using ...
The most common treatment of anisotropic friction in the literature assumes an ...... never outside the slip surface, therefore there are always two real solutions for ...