Page 675 - NGTU_paper_withoutVideo

Modern Geomatics Technologies and Applications

frequently affected by earthquakes in the recent years. This city is the most populous city in the northwestern of
Iran and is the economic and political hub of the northwestern Iran. The occurrence of any high-intensity
earthquake will cause large casualties in this city [4]. Therefore, the development of operational plans to deal with
crises and the training of aid workers are necessary.
Determining the number of rescuers is difficult due to the complexity of environment's crises. Simulation can
be a practical and operational tool to determine the number of aid workers. Natural disasters affect people's lives
[4]. Crisis environments are a complex system that is affected by various factors. Spatial-based simulation of
earthquake is one way to accurately assess the outcome of emergency plans [5]. Proper simulation of crisis
environments can provide a variety of information for managers and decision makers and is extremely helpful in
determining the number of aid workers. Quasi-realistic scenario in case of planning for quick responses is the
basis of earthquake-disaster management using decision-making techniques and is effective approach to
earthquake mitigation that provides opportunities to examine future events [6]. The simulations make
opportunities for communities to improve their understanding of earthquakes as well as their specific level of risk
[1]. The activities of search and rescue of victims in large-scale crises situations are highly concerned with social
problems [7], so that the activity working well in commercial circumstances may not be appropriate in handling
disasters [8]. One of the main concerns in complex systems is how to simulate them. Agent-based modeling
(ABM) is one of the practical tools in simulating complex socio-technical collaborative systems like the
earthquake. ABM is now widely accepted, especially in case that the problem's fields is particularly complex,
large, or unpredictable, and the system is distributed and open [9]. One of the most efficient ways of exploring
the problem is to use ABM [5, 10]. ABM with the bottom-up approach firstly separate the components of a system
and then analyze them. In this type of simulation, first the environment is simplified and different components of
the environment are identified. These components are then simulated in terms of capabilities, limitations, and
objectives of agents. These simulated agents start to operate in environment based on the defined characteristics
and their interaction with other agents [9]. In addition, the performance of the system can be examined by changing
agents' characteristics or environmental conditions, and new goals can be set.

Modeling of earthquake environments has been performed in various research. Saoud et al. (2006) described
a multi-agent-based approach for modeling the dynamics of large-scale disaster situations. They focused on two
organizational strategies and various heuristic algorithms in agents' optimization. Several analyses have been
simulated for each of the typical scenarios' configuration. Simulation results indicated that electronic
communication reduces delays and victims’ losses; however, centralized rescue process remains as the most
efficient in case of having many victims and few rescuers [5]. Farenilly et al. (2003) introduced a multi agent
system based on the RoboCup Rescue simulator to improve decision makers' understanding of earthquakes and
their specific level of risk. They described a framework for Cognitive Agent Development and a methodology for
evaluation of multi-agent systems that aims at measuring the efficiency of a system, and its robustness in uncertain
and complex environment [7]. Peng et al. (2014) developed dynamic environmental factors into the disaster-relief
supply chain and the impacts of road condition and delay in information transfer were modeled. They showed that
the road condition and the choice of the inventory planning strategies influence the system performance
significantly [8]. Due to the complexity of the environment, in some studies, only a part of the environment after
the earthquake (e.g., fires following earthquake) has been simulated [11]. Hooshangi and Alesheikh (2018)
simulated the urban search and rescue operation (USAR) by Multi Agent Systems (MAS). They simulated search,
rescue, and medicine agents in District 3 of Tehran (Capital of Iran), and developed an algorithm based on the
contract net protocol (CNP) method by considering interval uncertainty. In their study, the proposed method for
tasks allocation between agents was evaluated for different number of agents while the minimum number of rescue
agents was not calculated. Their main focus and purpose were examining their proposed method of assigning tasks
[12].

Determining the number of rescuers has been studied in limited research and there is no clear approach to
calculate it. Hassanzadeh et al. (2013) presented a GIS-based application by taking into account specific

2