Modelling and simulation of assisted hospital evacuation using fuzzy-reinforcement learning based modelling approach

Mohammad Abir, Intiaz and Mohd Ibrahim, Azhar and Toha, Siti Fauziah and Mohd Romlay, Muhammad Rabani (2024) Modelling and simulation of assisted hospital evacuation using fuzzy-reinforcement learning based modelling approach. Neural Computing and Applications. pp. 1-30. ISSN 0941-0643 E-ISSN 1433-3058

	PDF - Published Version Restricted to Repository staff only until 2 February 2026. Download (3MB) | Request a copy
	PDF - Published Version Restricted to Registered users only Download (201kB) | Request a copy

Abstract

Available hospital evacuation simulation models usually focus on the movement of the evacuees while ignoring the crucial behavioural factors of the evacuees’ which impact the simulation results. For instance, the issue of patient prioritization behaviour during evacuation simulation is often overlooked and oversimplified in these models. Furthermore, to control the movement of the evacuees, almost all these models utilize rule-based artificial intelligence to develop navigation systems, which sometimes do not guarantee realistic and optimal movement behaviour. This research aims to address these problems by modelling feasible and novel solutions. In this research, we propose to develop a hospital evacuation simulation model which utilizes a hybrid of fuzzy logic and reinforcement learning to simulate assisted hospital evacuation using the Unity3D game engine. We propose a novel and effective approach to model patient prioritization using a fuzzy logic controller; a reinforcement learning based navigation system to tackle the issues related to the rule-based navigation system by proposing novel reward formulation, observation formulation, action formulation and training procedure. The results and findings exhibited by the proposed model are found to be in line with the available literature. For instance, available literature suggests that an increased number of patients increases the evacuation time, and an increased number of staff or exits decreases the evacuation time. The proposed model also demonstrates similar findings. Moreover, the proposed navigation system is found to take a ‘‘near shortest distance’’ to reach the target as the mean difference between ‘‘shortest vector distance’’ and ‘‘distance covered’’ is approximately 1.73 m. The proposed simulation model simulates the repeated patient collection more realistically and can be used to estimate the Required Safe Egress Time, which enables the establishment of safety performance levels. The evacuation performance of different scenarios can be compared using the proposed model. This research can play a vital role in future developments of hospital evacuation simulation models.

Item Type:	Article (Journal)
Uncontrolled Keywords:	Assisted hospital evacuation Evacuation simulation Fuzzy logic controller Reinforcement learning
Subjects:	T Technology > T Technology (General)
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Engineering Kulliyyah of Engineering > Department of Mechatronics Engineering
Depositing User:	Dr Azhar Mohd Ibrahim
Date Deposited:	24 Jan 2024 14:50
Last Modified:	29 Jul 2024 11:39
URI:	http://irep.iium.edu.my/id/eprint/110540

Actions (login required)

View Item