Multi agent modelling and simulation

A. GAMA Platform Links

Catalogue of online materials - This is a list of some of the online materials (video, tutorials, pages) on the GAMA platform.
Code Examples - GAMA provides a rich set of code examples (written in GAML) that are available within the application. These examples can provide guidance on the syntax, and an excellent reference in developing and building models and simulations using GAMA. They can be used as building blocks to develop more complex models for different domains.

B. Other Platform Links

C. RESEARCH

A catalogue of some of my multiagent models and simulations developed using the GAMA geospatial multi agent modelling platform. The video and simulations are from October 2014 to the present.

These can be similarly found in my simulation YouTube Channel: https://www.youtube.com/channel/UC1-lMBRbHIYzLdu7YUy-laQ/videos.

Type	Video (HAL, YouTube)	Description
PorterrSIM, Multi Agent Modelling of Logistics Flows in the Seine Maritime Axis, The Case of the Automotive Sector Date: 2021-10-11, Location: https://hal.archives-ouvertes.fr/hal-03465141		The video shows/demonstrates the simulation developed in the paper by Julius Bañgate, Thibaut Démare, Dominique Fournier, Eric Sanlaville. Présentation d'un outil de simulation des Flux en vallée de Seine. International Conference on Smart Corridors and Logistics 2021 (ICoSCaL21), Nov 2021, Le Havre, France. ⟨hal-03465535⟩
Presentation: "Multi agent modelling of seismic crisis", This presents the model: SOLACE (SOciaL Attachment and Crisis Evacuations). In the video, the my session starts at 1:30:31 and ends at 2:15:00 Date: Nov 26, 2021		This presents SOLACE, the multi agent model developed during my PhD, in great detail. SOLACE is a geospatial, microscopic model and social simulation of human behaviour and pedestrian evacuation during seismic/earthquake crisis. The research considered social attachment is the main driver of differerent behaviours and social interactions during evacuation. Threat or crisis situations trigger proximity seeking behaviours, making individuals seek attachment figures/objects/places etc. that give comfort or solace during threat. The model also implements different earthquake intensity scenarios that cause damage to structures/building in the model. Debris are produced by damaged buildings in the simulation depending on the typology of the structure, its height, and the intensity of the earthquake. These debris result to: (a) blocked pathways trapping human agents and hindering evacuations; and/or (b) injuries or death to human agents who may be struck or entombed by debris. This was a plenary presentation at the 3 day PhilGEOS2021 Virtual Conference last November 18, 2021. The presentation starts at 1:30:33
SOLACE-Evacuation 1 Date: May 24, 2020, before Dec 18, 2019		Agent evacuation as implemented in SOLACE. This short demo video shows how microscopic multi-agent/pedestrian earthquake evacuation is implemented in SOLACE. Dynamic social interactions as well as interactions with the physical elements in the seismic crisis environment (debris/barriers/safe zones) are shown in the simulation. Pre-evacuation behaviours, and the location of agents in the building (floor level), affect the times of egress from structures, and eventually the arrivals in safe areas. Social attachment influences evacuation, and can result to following behaviours. The increasing level of building damage during the earthquake duration is shown as a gradient of colors from light grey, orange, dark red, to maroon. Debris blocks are created/deposited randomly around damaged buildings during the duration of the earthquake. Human agents can be injured or killed when hit by debris. Doors and pathways/roads/alleys can likewise be blocked hindering evacuation or trapping agents. Human agents react to the presence of debris, injured agents and casualties. SOLACE is developed from the GAMA multi-agent platform. Link to Phd Thesis: https://tel.archives-ouvertes.fr/tel-02613082, Link to slides: https://mamscgfr.weebly.com/slides---thesis-defense.html.
SOLACE-Evacuation 2 Date: May 24, 2020, before Dec 18, 2019		Agent evacuation as implemented in SOLACE. This short demo video shows how microscopic multi-agent/pedestrian earthquake evacuation is implemented in SOLACE. Dynamic social interactions as well as interactions with the physical elements in the seismic crisis environment (debris/barriers/safe zones) are shown in the simulation. Pre-evacuation behaviours, and the location of agents in the building (floor level), affect the times of egress from structures, and eventually the arrivals in safe areas. Social attachment influences evacuation, and can result to following behaviours. The increasing level of building damage during the earthquake duration is shown as a gradient of colors from light grey, orange, dark red, to maroon. Debris blocks are created/deposited randomly around damaged buildings during the duration of the earthquake. Human agents can be injured or killed when hit by debris. Doors and pathways/roads/alleys can likewise be blocked hindering evacuation or trapping agents. Human agents react to the presence of debris, injured agents and casualties.
SOLACE-Evacuation (Zoomed, Oblique) Date: May 24, 2020, before Dec 18, 2019		Agent dynamic social interactions, and interactions with the crisis environment, during evacuation. This short demo video shows how microscopic multi-agent/pedestrian earthquake evacuation is implemented in SOLACE. Dynamic social interactions as well as interactions with the physical elements in the seismic crisis environment (debris/barriers/safe zones) are shown in the simulation. This zoomed-in view shows how agents navigate in free space, how they perceive others (from social groups) and their environment (via perception angle/line/distance), and their resulting social interactions. The spatial constraints imposed by barriers (debris, buildings) to agent evacuation can likewise be seen. The emergent clustering of agents in safe areas is also shown. SOLACE is developed from the GAMA multi-agent platform.
SOLACE Demo 2-2 (Zoom, Top View) Date: May 24, 2020, before Dec 18, 2019		Agent dynamic social interactions, and interactions with the crisis environment, during evacuation. This short demo video shows how microscopic multi-agent/pedestrian earthquake evacuation is implemented in SOLACE. Dynamic social interactions as well as interactions with the physical elements in the seismic crisis environment (debris/barriers/safe zones) are shown in the simulation. This zoomed-in view shows how agents navigate in free space, how they perceive others (from social groups) and their environment (via perception angle/line/distance), and their resulting social interactions. The spatial constraints imposed by barriers (debris, buildings) to agent evacuation can likewise be seen. The emergent clustering of agents in safe areas is also shown. SOLACE is developed from the GAMA multi-agent platform.
3D Visualisation Date: Aug 6, 2018		Gama 1.8 3D Visualisation, Grenoble
Social Behaviour Date: Apr 24, 2017		Trying to simulate social interaction. Green lines define the focus of parent agents (Triangles) towards evacuation areas (Grey polygon), Red lines between parent agents and child (Circle) agents, Blue lines between parent agents. Child agent can only go with parent agents in the same social group (defined by colors: Blue, Green, Red and Yellow). In this simulation, parents from different social groups can interact.
Social Behaviour, Agents, Walls, Doors and Outdoors ;) Date: Apr 25, 2017		Testing agent interaction within simple building layouts. Interaction depending on context (indoors or outdoors) differ. Agents can exit through doors. More interaction with other agents occur (for this simulation) outside buildings. Now able to work with more spatial elements (detail) ;)
Social Behaviour_ Pick and GO Date: Apr 21, 2017		In this simulation, agents with the same ID (color) go together. Triangle agents (red, green, blue and yellow) move around buildings. When close enough, Circle agents (located in doorways) with the same ID (color) follow. A white line can be seen connecting the agents. This may simulate actions such as: (a) parents fetching kids from school, (b) shopping, (c) passenger pickup, ...etc. ;)
Social Behaviour_ Pick and GO with Chart Date: Apr 22, 2017		n this simulation, agents with the same ID (color) go together. Triangle agents (red, green, blue and yellow) move around buildings. When close enough, Circle agents (located in doorways) with the same ID (color) follow. A white line can be seen connecting the agents. This may simulate actions such as: (a) parents fetching kids from school, (b) shopping, (c) passenger pickup, ...etc. Charts (pie and series) display status for each agent group ;)
Large Agent (200K) Population Test with GAMA (displaying interface) -2 Date: Apr 18, 2017		Able to run faster evacuation simulations with vector free space and a large population of agents (in this case 200,000 individuals).
Debugging agent social behavior during evacuation and 3D visualization (GAMA 1.7) Date: Sep 17, 2016		Debugging agent social behavior during evacuation and 3D visualization. Debris (orange) now are randomly generated near building walls in the urban center. This will be improved in the next version to assign debris and damage probabilities based on building/material and type and earthquake intensity; also more psychological and social factors/dynamics ;) GAMA simulation. Best viewed in HD.
April122017 evacuation simulation test_Grid space. Date: Apr 12, 2017		GAMA evacuation simulation.Today's test. Many agents.
Evacuation BDI_1 Date: Feb 13, 2017		Evacuation simulation, Grenoble. Cited in https://news.cnrs.fr/articles/modeling-the-panic-moment
Evacuation BDI_2 Date: Feb 13, 2017		Evacuation simulation, Grenoble. Cited in https://news.cnrs.fr/articles/modeling-the-panic-moment
Evacuation BDI_3 Date: Sep 16, 2016		Evacuation simulation, Grenoble. Cited in https://news.cnrs.fr/articles/modeling-the-panic-moment
Indoor Hospital Evacuation 1 Date: Jun 18, 2016		Building Indoor Evacuation (GAMA Simulation) - 200 agents, egress at 2 ground level exits (stairwell) to a safe area (light green), several meters away from the building. [View in HD]
Indoor Hospital Evacuation 2 Date: Sep 6, 2016		GAMA Evacuation in a complex layout, testing non overlapping agents, OpenGL. Best viewed in HD.
Urban Mobility_Home_Work 1 Date: Oct 13, 2015		Road network and building data (shapefile) from OpenStreetMap. Variable work hours and defined work (grey buildings) and home locations (white buildings). Using Patrick Taillandier's code with a bit of modification. Simulations run smoother on a mac, and a bit coarse on Ubuntu. Best viewed as HD.
Urban Mobility_Home_Work 2 Date: Nov 12, 2014		Grey polygons are dormitories (homes) and the empty ones, work places. Generated 500 agents who travel by car. Travel is made from home-work at defined time intervals. Charts show traffic and distribution of people at work and home every 10 time steps. Shapefiles are from Openstreetmap, edited with QGIS. (Best viewed on HD)
Urban Mobility_Home_Work 3 Date: Nov 12, 2014		Grey polygons are dormitories (homes) and the empty ones, work places. Generated 500 agents who travel by car. Travel is made from home-work at defined time intervals. Charts show traffic and distribution of people at work and home every 10 time steps. Shapefiles are from Openstreetmap, edited with QGIS. (Best viewed on HD)
Urban Mobility_Home_Work_Traffic Date: Oct 13, 2015		Road network and building data from OpenStreetMap. Best viewed as HD.
UP Diliman Campus Simulation (3D, color) Date: Oct 26, 2014		UP Diliman Campus Simulation (3D, color). "Walking on campus" powered by GAMA and OpenStreetMap data. Best viewed on HD
Epidemiology_Infection Spread 1 Date: Jun 15, 2016		GAMA Simulation ;) - Infection spread (e.g. Flu) in a small urban center - 5 initial agents infected; Population: 10,000 agents; Infection distance=5 meters; Probability of infection=50%. Agents move building to building via roads/pathways at hours 9h (go to work), 12h (lunch time) and 18h (go home) - RED - Infected, GREEN - Susceptible; (VIEW IN HD); reference: http://vps226121.ovh.net/tutorials#In...
Epidemiology_Insect bite (e.g. dengue) Date: Jun 17, 2016		Modeling Infection Spread via Insect bite (e.g. Mosquito-Dengue, Zika, other) in a Small Urban Center (GAMA Simulation) - insects (small circles) move randomly; people (large circles) move from building to building @ 5 km/hr via road network 9am (go to work), 12n (have lunch) and 6pm (go home), RED-Infected, GREEN - Susceptible; people get infected (50% probability) when bit by an infected insect; a susceptible insect gets infected (50% probability) after biting an infected person. [View in HD]
Crowd Dynamics, Evac sim with GAMA 1.7 - Leader-Follower, Lane Formation, Arching and Clogging Date: Sep 8, 2016		Evacuation simulation with GAMA 1.7 - Leader-Follower behavior with Lane Formation, Arching and Clogging @ passageway and evacuation area, using improved code and larger leader influence (video @ 30x simulation speed). Best viewed in HD.
Crowd Dynamics with GAMA 1.7 - Arching and Clogging ;) Date: Sep 6, 2016		Crowd Dynamics with GAMA 1.7 - Arching and Clogging. View in HD ;)
Modelling Happiness, Residential Segregation - 90% Corruption, Allowed to Build Date: Jan 21, 2015, GAMA Summer School		Modeling the effects of happiness, taxation and corruption (0.9%) on residential segregation and the formation of informal settlements using GAMA. From last week's GAMA Training @ UPD, Group 2. Codifying behavior and policy. Seeing emergent properties (clustering, new structures) as agents navigate through geographic space and time. This model was developed by my group during the GAMA summer school held in the Philippines in January 2015 by the GAMA Development team. This model really drove the point and showed the power/benefit of integegrating abstract but felt societal concepts (e.g. happiness, culture, social issues), spatial data, ideas, policy, multidisciplinary collaboration and computation in multi agent models and simulation