Plusicus simwateuns su- ciaux & base d’agents ont été développés pour comprendre des comportements humains cL prévoir Ja situalion juste apres une catastrophe voir le ehapitre 2 Pest la,
Comportements des agents 2 2.2
Comportement de perception des obstacles 28
L’agent Human peut observer des obstacles et des mes bloquées dans sa zone de per- ception, mé:oriser les rues bluquées observes pour éviler & Favenir, 8l entre dang les
Chapler 5 Conception du mudéte 29 ằ pa ng XK NO
T9MEHR CỦ, i rLiHeopds point oN
In areas with small and medium obstacles, movement is hindered, causing slower progress due to various debris When navigating through zones with larger obstacles, the likelihood of fatality increases significantly.
To avoid significant obstacles, the idea is for the agent to find an alternative route to reach their target, whether it be with leaders or by navigating to a new goal with wandering agents The simplest method is to update knowledge of blocked streets and the path to the target when the agent encounters an obstacle However, this approach can lead to problems.
Dans le modéle, on obscrve certaines configurations de rues bloquées (figure 5.3}
The red circle represents a significant obstacle, while the blue square symbolizes the agent, Human The agent encounters a situation where there is no clear path to navigate The segment of the street where the agent moves before encountering the obstacle is longer, whereas the segment of the street before reaching the obstacle is shorter The agent is positioned between two ends of the street where obstacles are present Additionally, the agent can see the obstacle on another street, as it is in close proximity.
Chapter 5, Coneeption dụ moièke 30 xi ứ đ o
FIGURE 5.3: Configurations des rues bloquées
In configuration (2), the agent finds itself on a blocked street and must turn back to take an alternate route However, the segment required for the turnaround is longer than the remaining length of the street, leading the agent to frequently opt for entering the obstacle To address this issue, the agent should return to the previous intersection after encountering an obstacle The weights of the blocked streets are then considered.
In 4 days, leaders opt for alternative routes to reach their targets, while agents select different objectives to pursue If an agent encounters an obstacle in configuration 5, they typically update the points of interest (POI).
FIGURE 5.4: Agent se déplace vers le point dintersection
Deuxitmement, pour les configurations (1) et (4), agent n’a pas de choix, il passe de obstacles pour sortir Afin de résoudre ce probléme, on ajoute un paramétre concernant
The maximum number of times the agent encounters obstacles is 2 to 3; after this, the agent stops and remains in place This addition is suitable for 4 psychological aspects of individuals during evacuation, as they attempt to find a path to their destination but ultimately stop Figure 5.5 illustrates the concept of the agent's destination, highlighting the behavior of obstacle perception within a simulation cycle.
Chapler 5 Conception du mudéte aL tac anon a we eed vue) a
—_ et Ls sat ns te ta nti ne ni m ` ` es
Tmapora_mreaccn > stem ones knoneage
FIGURE 5.5: Processus de perception de grands obstacles
In the initial phase, followers and cautious agents randomly search for a leader among themselves Once they identify a leader, they begin to follow and adapt their targets, updating their position and the paths to pursue the leader If they fail to find a leader, they wander and persist in their search Figure 5.f illustrates the diminishing process within a simulation cycle.
Leaders adjust their speeds to accommodate their followers during the 10 simulation cycles This approach is essential, as a leader cannot constantly seek out their followers, and doing so would be inefficient.
Chapler 5 Conception du mudéte 3 temps de calen| [ut 10 cycles de simulation est siftisant, pour que ie leader puis leurs suiveurs La igure 5.7 uiontre le processus de réglage de là vitesse
(am) uwrce-eotowers LZ NC ie sneec > slowest see
Friacre 5.7: Processus de réglage de Ta vilesse
There are two behaviors observed for the agent interacting with obstacles: street damage and human damage The agent seeks to block or destroy any overlapping streets or humans The street damage behavior occurs once during the simulation at the initialization step, as there are no changes to the obstacle and street agents during the simulation In contrast, the human damage behavior is executed at each iteration of the simulation.
Parameters are categorized into Picbier GIS, General Human Agent, and Human and Shock Communication (label 5.2) By varying the values of these parameters, we can generate different simulation scenarios that can help analyze various information campaigns For instance, to assess the impact on the number of casualties during a nighttime earthquake, we can assign a value of 100 to the parameter "Probability of persons in buildings," indicating the assumption that everyone is at home during the night Additionally, there are remarks regarding the probability of vehicles.
Fichier shapefile pour rues Iiehier GI Wichier shapefile pour batiments
Fichier shapefile pour espace vert
Agent Ituman | Prohabilité de femme 0,100), đoat
: Probabilité de gens avec “age < 40" [0,100], float lùng Probabilité de gens avec “40