Multi-Agent Geo-Simulation in Support of Risk Management by zxq20493


									           Serious Games and
       Multi-Agent Geo-Simulation
          for Decision Support
2nd International Workshop on Mobile Geospatial Augmented Reality
                     Quebec city August 2008

                             Dr. Bernard Moulin,
                               Laval University,
           Computer Science and Software Engineering Department,
             Pouliot Building, Ste Foy Quebec G1V 0A6, Canada
                Tel. (1) 418 656 5580 Fax (1) 418 656 2324
Games are becoming ‘serious business’ …
Digital Games vs. Games for Learning
Market Opportunities for Serious Games ?
The Development of SG can take advantage of research …
From Risk Anticipation to Risk Management
Multi-Actor Dynamic Spatial Situations (MADSS)
Decision Making and MADSSs
Multi-Agent Geo-Simulation (MAGS)
Samples of Past Project in MAGS (at Laval University)
   Crowd Simulation using the MAGS Platform; Simulation of Shoppers’ Behaviours in Malls
   Fire-MAGS; VNO-MAGS Project; COLMAS Project
Serious Games might benefit from MAGS research
Proposal for an Architecture Integrated Simulation and Gaming Modules
MAGS for Decision Support: The scientific side
Synergy between MAGS and Serious Games ?                                                   2
Games are becoming ‘serious business’ …
 … in various domains such as health care training, emergency preparedness
 and incident management training
 Using game technology for non-entertainment applications → game-like
 activities with a serious purpose
 “The Serious Games Initiative is focused on the uses of games in
 exploring management and leadership challenges facing the public
 sector …use of games in education, training, health, and public policy”
 Game-Based Learning (GBL) is promising because it is well known that
 entertainment games are ‘engaging’ and have motivational virtues that
 typical learning system lack
 Games are particularly appropriate for training people to better function as
 members of teams in complex settings
 Multi-player games are excellent vehicles for helping participants to better
 coordinate and cooperate: they are used to train individuals and teams in
 diverse industries                                                           3
Digital Games vs. Games for Learning
    Table 1: Differences between digital games and games for learning in purpose, rules,
                        play and culture (

                           Digital Games                               (Serious) Games for Learning

Purpose   For entertainment purpose. Context presented are       For learning and skill development purposes.
               mostly fictitious or fantasy based.                    May be a form of entertainment based on
                                                                      the interpretation of user.
 Rules    Rules are designed to accommodate the activity of      Rules are designed for specific learning
              play which are often be tuned for playability          outcomes that can be used to measure the
              rather than reflecting the real happening.             interactions during the “serious play”.
                                                                     Rules can be simplified or made complex
                                                                     to support the activity of play.
 Play     Interactions primarily designed for entertainment      Interaction designed for learning purpose with
               purposes with directed objectives which can            meaningful responses and measurable
               be driven by storytelling. Interactions                outcomes. Knowledge is disseminated
               resemble the real world interaction in a               through events triggered by optimised
               simplified or abstract approach.                       interactions and dialogues.
Culture   Beliefs, norms and world setting presented visually    Beliefs, norms and world setting presented
               and via narrative are often set in an imaginary        visually and via narratives should be set in
               world which is represented artistically and            a real setting to reflect truthfulness and
               exaggeratedly.                                                                                     4
                                                                      have direct and explicit relation to the real
                                                                      world events.
Market Opportunities for Serious Games ?
(L. Jackson DTI 2007)
  Market size for serious games?
     Worldwide CG market £25 billion+ (Screen Digest/ELSPA 2006)
     E-learning market estimated at more than US$80 billion (ITI Techmedia)
     Serious games in USA estimated at $50 million (SGI, GDC 2006)
     Potential market for serious games US$700 million (ITI Techmedia)
  Industry/academia spotting the market opportunities:
     Annual Serious Games Summits in Washington attract major US & other
     government commissioners
     Nintendo launches Dr Kawashima’s Brain Training in Europe in June 2006; in
     UK top ten best-seller list for months (ELSPA ChartTrack, April 2007); Sony
     launches PSP in Education Initiative (March 2007); also PS3 “Home” (March
     ANGILS, first European networking trade association for serious games,
     formed in late 2006
     The Serious Games Summit GDC (San Francisco 2008)
     A growing number of companies are active in the Serious Game area
The development of SG can take advantage of research …

  … carried out in various domains such as:
     the long tradition of story telling,
     the scenario-oriented role-playing games used for example as a method
     of education in business and management schools …
  but also in the simulation field (i.e. flight simulators) …
  Yet, other less known research domains may contribute to the
  development of advanced and scientifically grounded serious
  games such as those related to geo-sciences and artificial
  intelligence, and in particular multi-agent geo-simulation
  Let us look at ‘serious subjects’ and assess their potential for
  the development of serious games

  From Risk
   to Risk
   The need to
   understand /
 assess / simulate
complex situations

  S.L. Cutter
  Vulnerability to
  hazards, Progress
  in Human
  Geography, vol.
  20 (4), pp. 529-
  539                 7
Multi-Actor Dynamic Spatial Situations (MADSS)
  MADSSs involve a large number of actors of different types (human,
  animal,…) acting in geographic spaces of various extents
  MADSSs need to be monitored to insure :
     human security and equipment preservation in case of natural or man-
     provoked hazards (flood, earthquake, wildfire, oil slicks),
     the respect of public order (population evacuation, crowd monitoring and
     control, peace-keeping, etc.)
     the adequate use of infrastructures (monitoring of people and households
     transportation and shopping habits in a urban area to better plan transportation
     infrastructures, location of services’ retailers, etc.)
     Impact of emergency response plans
  Decision makers need an overall understanding of the
  situation to monitor its evolution, to develop strategies to
  adequately intervene, to develop and compare alternative
  intervention scenarios and to anticipate the consequences of
  these interventions
Decision Making and MADSSs
 Most critical MADSSs involve large populations and/or
 important infrastructures/equipments that are usually spread
 out on large geographic spaces
 Natural hazards or man-provoked hazards can trigger
 MADSSs that may endanger human lifes, threaten the
 equilibrium of eco-systems and/or the preservation of
 equipment & infrastructures
 Decision makers need tools to simulate such MADSSs in order
 to anticipate how they may evolve and to possibly assess the
 impacts of their decisions
 Systems based on Multi-Agent Geo-Simulation (MAGS) may
 provide some support to decision makers, either by training
 them before the occurrence of such situations or by providing
 them with simulations during the evolving MADSS
Multi-Agent Geo-Simulation (MAGS)
 ‘Geosimulation deals with the construction of different kinds of
 spatial models in order to study spatial phenomena while
 developing software to support ‘actor-based’ simulations
 Multi-Agent Geo-Simulation (MAGS) takes advantage of the
 coupling of Multi-Agent-Based Simulation (MABS) and
 Geographic Information Systems (GIS).
 Multiagent geosimulation can be effectively used to simulate
 complex systems in virtual georeferenced environments.
 MAGS tools/approaches used to support decision makers for:
    the creation of action plans and the assessment of these plans and of their
    impacts on actors’ mobility and decision making,
    The comparison of intervention scenarios
    to help visualize the outcomes of action plans
Crowd Simulation Using the MAGS Platform (1)

                              Simulating at a micro level
                              the mobility behaviors of

Crowd Simulation Using the MAGS Platform (2)


  Simulation of Shoppers’ Behaviours in Malls

Comparison of scenarios
modifying the physical
configuration of the mall and
assessment of their impact on              13

shoppers’ mobility
           The ENCASMA Approach
   and Geosimulation of Control of Forest Fires

              Pathfinder réel                         Agent Pathfinder

                   Aspects                              Modèles pour
               dynamiques de                             simuler les
              l’environnement                           aspects dyn.
                                                       Plate-forme de
                Espace réel             GIS              simulation

               Real World                           SimulatedWorld

                         In such simulations several issues
                         are related to mobility problems, as
                         for example for the evacuation of
                         endangered populations

       We worked on the simulation of action
       plans for dozers to create fire breaks

                      VNO-MAGS Project

We simulate the dynamics of
the populations of Culex
mosquitoes and of crows and
their interactions which favor
the West Nile Virus infection

                                  Helping decision makers
                                  explore different intervention
                                  scenarios under various
                                  weather conditions
                                  (temperature, rain falls)
                                The COLMAS Project
       (Nsim Tech: Perron, Hogan; RDDC: Berger, Bélanger)

   •    Use of Multi-Agent Geosimulation and Machine learning for
        distributed continous planning in the case of territory surveillance
        using Unmaned Autonomous Vehicules (UAVs)
   •    Identification of the best displacement patterns that enable a group of
        UAVs to visit a set of targets

Various solutions found from the best one to the     UAVs’ patterns for scenario
        worst one (SDKsim –COLMAS)                         with obstacles          17
Serious Games might benefit from MAGS research

 MAGS projects can provide scientifically grounded and tested models to SG
 developers -> SG-MAGS joint projects
 MAGS projects are often used to simulate and analyse complex situations and
 to provide tools to decision makers to assess situations (MADSSs), to assess
 different intervention scenarios, to compare them, etc.
 The underlying models could be used to create a new generation of SGs which
 simulate such MADSSs, enabling players to intervene in the game as the real-
 life decision makers:
    Example1: our MAGS application involving crowd’s and control forces’ behaviors
    and interaction, using non-lethal weapons, could be used to create a serious game to
    train police force commanders
    Example2: our VNO MAGS application could be used as a starting point to create a
    Serious Game for public health officers who monitor zoonotic diseases
 In some MAGS Projects, lot’s of real data are collected from various sources,
 processed and transformed in order to feed the geosimulations, to test various
 models, etc.: This kind of data might also be useful for Serious Games
Proposal for an Architecture Integrating
Simulation and Gaming Modules (Jain and McLean 2005)
1) Creation and management of a federation of simulation and gaming
    modules appropriate to represent the selected incident management
2) Integration of heterogeneous simulation federates modeling different
    interrelated aspects of the emergency event;
3) Integration of heterogeneous gaming modules with trainees role-playing
    within the same locale in the emergency event simulation;
4) Synchronization between the macro-level modeling of simulation and
    micro-level modeling of gaming modules;
5) Control over execution of both simulation and gaming modules through a
    training manager console;
6) Execution in Massively Multi-player Online Games (MMOG) mode to
    support a large multi-agency incident management exercise;
7) Access to heterogeneous data servers for supporting simulation and gaming
8) Management of MMOG execution.
Gaming sub-systems and Modules
(Jain and McLean: Integrated simulation and gaming architecture for incident management training, Winter
Simulation Conference 2005)

 MAGS for Decision Support: The Scientific Side
                                                      models                    Spatio-temporal
                                                                                  Analyses of               Models
                                 Collect data                                    various kinds            (theories)
    World of                     About studied
     Interest                    phenomenon
                                                                                                          actors and
        +                                              Mobility Data                     Mobility         behaviours
     Studied                      Collect data         Population Data                   patterns
                                 about actors          Environmt. data                  and other
   Phenomeno                                                                              results
                                  Collect data                           Predictive
                                   about the                               models                         Model Actors
                                                                                                           and their
Interventions                            Generate information
                    Model the                                             Population attributes
                                              for MAGS
                  World of interest                                      Environment attributes

                                        Contextual                                      Models of
                      Specify           World Model
                                                                  Multi-Agent           actors and
                     scenarios                                                          behaviours
  Decision                                                           Geo-
  Maker (s)                               Scenarios               Simulation

                                    Present                                                   Calibrate
      Mission                      Simulation                                                Simulation        Model
     Objectives                   Results and                      Simulation
                                                                    outputs                  parameters         21
      Critical                    comparisons
Synergy between MAGS and Serious Games ?
 Importance of collecting plausible and complete enough data sets
 to feed the models and to calibrate and validate the simulations
 In the case of serious games developed in synergy with MAGS
 systems, researchers might take advantage of the deployed
 games and of simple tools to collect data on the players’
 platforms (it will be possible to assemble large amounts of
 experimental data in context)
 Getting such detailed and contextualized data would enable
 researchers to analyse sociological phenomena at deeper levels
 of inquiry and consequently to improve their models
 Such improvements would in turn be used to enhance the games
 that will be based on them.
 Many issues to be discussed … but also such a great potential
 lies in the complementarity of MAGS and Serious Games !!! 22
A nine Compartments Model for Situation Analysis, Decision
Support and Coordinated Intervention in MADSSs

        C9: Outcomes
              /              C2:             C3: Stat./
          Impacts         Observation         Spatial

                          C1: Situation   C4: Decision
                            (Real or
                         simulated in a
                         serious game)

             C7:                              C5:
       Dissemination /                    Anticipation
         Awareness           C6:            (MAGS)
           Building      Intervention


To top