Modeling & Simulating the Evacuation of a Building Based on

Building Floor Plan and Evacuation Strategies

Shreya and K.S. Rajan

Lab for Spatial Informatics, IIIT Hyderabad, Hyderabad, Telangana-500032, India

Keywords: Building Evacuation, Evacuation Strategies, Floor Plan, Path Planning.

Abstract: Accidental fires in public and large buildings not only cause property loss but also can lead to loss of lives.

During such emergencies, building evacuation depends on a range of factors including floor plans, exits

available, obstructions if any, the occupancy levels of the building, and so on. The study here brings together

the spatial, temporal, and path planning possibilities to evaluate fire evacuation strategies for 2D building

plans. It provides a geospatial framework to assess the impacts of dynamic changes in the building

environment and its impact on evacuation outcomes. In this study, occupancy-based path planning using

Pgrouting over an IndoorGML formatted data is combined with modeling their interactions over the path

toward the exit to assess the outcomes. This computational approach over the time-dependent path provides

interesting insights into determining the number of paths and the need for one or more exits during an

emergency. The study shows that integrating the floor plan into path generation and people flow can be a

powerful tool for assessing the building environment.

1 INTRODUCTION

Evacuation strategies are important in many

situations even in regions that may not be affected by

earthquakes or other natural causes. As urban areas

get more densely populated with multistoried

buildings that have many people either living or

working from these, it is important to evaluate the

evacuation strategies in these buildings. Engineers

and Architects need to demonstrate that their designs

of the building are safe for the residents. While they

may be also concerned that providing multiple exits

can affect their designs, a better understanding of

safety can help improve the designs suitably. On the

other hand, a geospatial model of the indoor floor

plans can help improve the understanding by not only

visualizing but allowing for multiple computational

models to be evaluated using geospatial models of the

evacuation paths. Indoor navigation paths have been

derived, in recent years, based on floor plans (Yang

and Worboys, 2015) and have also been extended to

data models like CityGML or IFC ((Kolbe et al.,

2005). While these approaches provide for a path

from say a room to the building exit, they are

generally independent static paths for each person

assuming no crowding or delay in movement when

all the occupants rush to the exit in an emergency.

In an emergency, a path based on quantitative

measurements like distance, direction, and angles is

not of much use or can be rather difficult to

comprehend by people trying to move in a hurry. If

these spaces can be captured qualitatively and shared

as part of the evacuation path to the users, it might

make it easier for people movement and re-defining

the paths depending on the changing environment in

the building. A semantic indoor space model

proposed in (Maheshwari and Rajan, 2016) based on

combining the ontology for indoor spaces with

geometric and semantic characteristics of the space,

as defined in (Maheshwari et al., 2019), can be

combined with the path generation approach to

provide for a more informative and comprehendible

path, that can be used by all residents of such

buildings. An attempt at using semantic information

in indoor path planning is presented in (Xiong et al.,

2015) by combining both geometric and semantic

information of building components.

Path planning in indoor spaces has been studied

based on shortest path (Ramón et al., 2013; Botsis and

Panagiotopoulos, 2021; Clementini and Pagliaro,

2020) algorithms like Dijkstra’s and Bellman-Ford’s

algorithms. While these shortest path algorithms

230

Shreya, . and Rajan, K.

Modeling Simulating the Evacuation of a Building Based on Building Floor Plan and Evacuation Strategies.

DOI: 10.5220/0011992600003473

In Proceedings of the 9th International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2023), pages 230-236

ISBN: 978-989-758-649-1; ISSN: 2184-500X

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

provide a good idea of the nearest exit, the optimality

was considered based on different criteria like least

risk (Mirahadi and McCabe, 2019), the best-known

path to the evacuee from among paths to multiple

exits (Liu et al., 2016) in a building. Geospatial

technology with the computational algorithms to

develop and verify whether a building satisfies the

emergency regulations is presented in (Ramón et al.,

2013). Indoor navigation approaches, as in (Mortari

et al., 2014) use displacement techniques to generate

routes and may be extendable for evacuation purposes

as they do contain topological information. Paper

(Sun and Liu, 2011) introduced a continuous

framework that talks about both structure and

topology, using a grid graph–based modeling for path

planning.

On the other hand, the availability of BIM models

of the buildings has led to exploring these datasets for

indoor path generation automatically based on

straight skeletons (Fu and Liu, 2019) or generating a

geometric topology network from the building spaces

and spatial connections as represented in an IFC file

(Taneja et al., 2011). In (Yenumula et al., 2015), they

used the BIM-based signage information to indicate

the exits that can be accessible during fires by

assessing the effect of smoke on the visibility of the

signages, while (Liu and Zlatanova, 2011) presents a

door-to-door approach and (Ma et al., 2017) talks of

using BIM over a schematic based evacuation plan

for the floors.

While all these studies do work on different

aspects of evacuation paths, their computation, and

representation, these are mostly static paths and do

not consider the changes in the indoor environment

like whether the exits are useable or reachable from a

given location in the building. This needs to be

evaluated as heavily crowded or occupied buildings

may need multiple exits for evacuation of the

residents in the shortest time. Also, the changes can

affect the evacuation models by affecting the time

along the path considering lags in people flow and

capacity of the paths in addition to the availability of

additional exits if any. This paper attempts to address

some of these issues in the generation of the optimal

path for evacuation for each occupant of the building

for a given 2D floor plan considering that the graph

network connecting these spaces may change in

adverse conditions.

1.1 Objective

In this study, an attempt has been made to integrate the

inputs from the floor plan generated network with an

option of new exits emerging in addition to the main

exits as part of an evacuation strategy to assess –

a) How the capacity of the path links or edges

can delay the evacuation time based on the

people’s movement pattern

b) How to determine if an evacuee will use an

alternative new emerging exit like a window

This paper assumes the availability of a space

model from IndoorGML, whose description is out of

the scope of this paper and develops on the possible

evacuation networks that emerge from such data.

2 METHODOLOGY

Figure 1: An overview of the methodology and its

components.

2.1 Building Plan for Node Generation

A building floor plan in Autodesk Revit Architecture

software is created with attribute information such as

door windows. The structural element of the building

contains six spaces denoted as a room, twelve

windows, and seven doors including one main exit.

All the rooms are accessible from one space to

another. This building plan, originally in DXF file

format, was imported into QGIS. This DXF file was

preprocessed in terms of topological consistency so

that it has the right number of polygon geometries,

location of doors, windows, etc. The building floor

plan was adjusted in QGIS according to the

coordinate reference system so that it lies on the same

plan of projection. Nodes were generated at the

centroid of the room, representative points for the

occupants of the room, and door center points for

each room space. In addition, to account for more

than one exit from the building, including emergency

or alternate exits like a window, the final destination

for the simulations is a node called the Evacuation

point away from the Building that is reachable from

Modeling Simulating the Evacuation of a Building Based on Building Floor Plan and Evacuation Strategies

231

all the exits, as shown in figure 3. These nodes are

further used for graph generation.

2.2 Primary Graph Network

The generated nodes are identified as the potential

sources, intermediate nodes, and destination nodes

such as main exits and windows as alternate exits.

The graph is then generated using the source and

destination pair of coordinates assigned to form the

primary network layer. The generated graph network

G can be seen in figure 3. Each path link is

characterized by its actual edge length and used in

calculating the path edge capacity.

2.3 Occupancy and Path Edge

Capacity

As the approach here is to estimate the impact on the

people flow movement in the context of an

emergency, it is assumed that the spaces (or rooms)

are fully occupied to their maximum capacity. Hence,

the occupancy of each space is estimated based on the

unit area needed per person. For the path edge

capacity, an average speed of 6 kmph is assumed per

person to account for the movement behavior during

evacuations. This allows for computing each

occupant’s travel time to the exit and hence

estimating the last man’s exit time. While the case of

emerging exits is like a window, a hold time is also

added to the link concerned to account for the delays

in accessing such exits. The proposed approach also

allows for a pass-through capacity to be added at the

specific nodes for estimating the rate of flow of

people.

2.4 Evacuation Path Generation

2.4.1 Shortest Path for the Base Scenario

In this study, a base scenario is defined as one that has

no obstruction and allows for the free movement of

people through the paths. This is possible from all the

structural elements such as doors and windows. Path

computation for the shortest path calculation to the

main exit is obtained using the Pgrouting module in

PostgreSQL. Here, the room spaces are considered

source nodes and the doors or windows on the

building periphery as the exits. Using the Dijkstra

algorithm, all potential shortest paths for each source-

destination combination are obtained, denoted by set

P. Each path, ( P

i,k

ε P ) denotes the i paths available

to each user or individual k. The shortest path for a

given user k is P

mk,

where m ε i, denotes the final path

chosen.

2.4.2 Evacuation Path for the Adverse

Scenario

In case of an emergency, the shortest path computed

in the earlier sub-section may be not available to the

user. Here, we consider that an untoward incident like

fire or another emerging condition can lead to a node

or an edge blockage. While the later disconnects only

the path that consists of the blocked edge, an affected

node disconnects all the edges that are connected to

it, as shown in figure 2. In this paper, the case of node

blockage is considered as the edge case can be a

special case of the same. The set P* is the updated

available set under changed circumstances. An agent-

based approach is used to evaluate the options from

the set P* and decide on the optimal path to take

considering the existing and alternate exits available

and the parameters computed earlier for the graph and

its link edges including lag due to link capacity. The

agent aims to minimize the time taken to the final exit

point.

Figure 2: Affected node and edge.

3 DATA

3.1 Floor Plan

A building floor plan covering a total carpet area of

102.349 m

as shown in Figure 3 is used for this

study. The building floor plan was converted to

shapefile format from DXF. It consists of six rooms,

six internal doors, one external main exit, and twelve

windows which act as an alternate exit during the

adverse case. The exit door has the capacity to carry

two occupants at a time whereas the alternate exit has

the capacity of one occupant at a time. In this study,

the number of spaces is taken as rooms denoted by its

centroids as the source points. Figure 3 shows the

floor plan followed by notations such as nodes that

are used for source and destination input, evacuation

point, the evaluation or safe point for evacuees to

reach, and the generated graph network. Network

graph notations are described in section 4.1.

Affected node

Affected Edge

GISTAM 2023 - 9th International Conference on Geographical Information Systems Theory, Applications and Management

232

Figure. 3: Building Structural Plan.

Table 1: Occupancy count of room.

Number of occupants based on the floor Space Model

Room 1 Room 2 Room 3 Room 4 Room 5 Room 6 Total

20 6 15 15 5 28 89

3.2 Occupancy

The occupancy is derived from the area of the room

space, as the model is evaluated for a worst-case

scenario of maximum occupancy. Table 1 shows the

occupancy in each space. There is no occupancy in

the corridor and the shaded room. It has to be noted

that occupancy is a key parameter here, and if it is less

than the capacity of the room it will affect the path

optimization and can lead to a different outcome.

4 RESULTS

4.1 Network Generation Including

Edge/Path Capacity

In figure 3, the line in different colors indicates the

generated graph. This graph accounts for the main

exit and the alternate exits and shows the shortest

paths only. Centroid-to-door distance is marked in

green, the centroid to the window is marked with red,

and the door-to-door distance is marked with purple.

These networks are assigned labels, indicating the

distance—time—capacity of the edge. The capacity of

an edge is proportional to the rate of people flow at

the exits. There are some limitations to people flow

based on the capacity - if the number of occupants is

greater than the least capacity edge of the path P

then not all the occupants can be evacuated in the

shortest time shown.

4.2 Path Identification

Path identification is shown in figures 4 and 5 which

have some representative agents for every room and

are uniquely equal to the number of potential exits for

a given space. Agents can be broad of two types – a

Modeling Simulating the Evacuation of a Building Based on Building Floor Plan and Evacuation Strategies

233

normal agent who uses the known path (P

) to the

main exit only; and a special agent who uses one of

the alternate paths. These special agents are activated

when the traditional or known path (P

) is disturbed

and their behavior is dependent on the conditions.

There are multiple evacuation agents in each space.

Each agent uses the graph or sub-graph appropriately

for the choice of exit and the path to be followed.

4.2.1 Base Scenario

The base case scenario is the free flow for exiting the

occupants with no blockage in between the path. For

example, room 1 has an occupancy of 20 and the

capacity of an edge-to-door is 16, which means that

the edge can hold only 16 occupants at a time hence

there will be a lag. If lag was not there, then it could

have been just the addition of all the time from source

to destination. If there was no lag it could have taken

only 22 seconds for an occupant to come out of the

main door instead of 254 seconds for the last position.

The base scenario is shown in figure 4 and the

evacuees are using the main exit as alternate exits will

take longer evacuation times.

4.2.2 Adverse Scenario

An adverse case scenario of fire in the building is

considered. The fire blocks the path to the exit and to

the evacuation point. The implemented scenario

shows the blockage at the most critical point, where

most of the edges are merging, representing the

worst-case scenario

For this scenario, as the node is

affected all the connected edges are dropped resulting

in multiple graphs being generated instead of one

main graph. In this case, there are four graphs

connecting alternate exits for room 1, room 4, and

room 5 along with one larger sub-graph for room 2,

room 3, and room 6 towards the main exit. The

alternate exits, and windows, are chosen by the agents

as these room spaces are disconnected from the

primary graph network. The evacuation time of the

last person was at 3195 seconds, comparatively quite

large from the base scenario. The movement of the

agents is depicted in figure 5.

4.2.3 Cumulative Count

In the base case scenario, evacuees took minimum

time for the exit as there is no lag or time delay for

opening/breaking exits, whereas in the adverse case,

scenario evacuees took maximum time due to time

delay/lag in breaking/opening the window and

coming out of it. Figure 5 shows cumulative count

graph results with a cascading effect that occurred

during the evacuation. The linearity of the graph

indicates that evacuees took minimum time for the

exit as there is minimal to no lag or time delay and the

non-linearity of the graph shows that evacuees took

maximum time due to time delay/lag in

breaking/opening the window while exiting.

5 CONCLUSION

The study proposed an effective methodology for the

evacuation modeling of the occupants based on the

changing paths to the exit based on a 2D structural

plan. Compared to the base scenario, the adverse

scenario took more than 12x times for evacuation of

the 89 occupants of this building. This indicates that

changed dynamics can have very different outcomes

and the provision of alternative exits with decent

capacity needs to be planned well. Also, it should be

noted that the capacity of the main exit and the access

time at the alternate exits affects the evacuation time

hence there is a need to evaluate these for different

cases and define the appropriate capacity of these

doors and windows. The use of an agent-based model

Base Case Scenario: No Fire Blockage

Figure 4: Base case scenario.

GISTAM 2023 - 9th International Conference on Geographical Information Systems Theory, Applications and Management

234

Adverse Case Scenario: Fire Blockage at a critical point

Figure 5: Adverse Case Scenario.

Figure 6: Graph: Cumulative Count.

to evaluate options provides for dynamic decision-

making in such situations or scenarios and adapts to

the changing graph networks. The adoption of a space

model fully can further help define the pass-through

spaces, occupied spaces, and their respective

constraints in a better way. Also, with the increasing

use of location sensors, it will be good to integrate

these models with a real-time people positional model

and crowding behavior to see how the scenarios will

evolve in varying ground conditions.

REFERENCES

N. Maheshwari, and K.S. Rajan., “A semantic model to

define indoor space in the context of an emergency

evacuation,” Int. Arch. Photogramm. Remote Sens.

Spat. Inf. Sci.—ISPRS Arch. 2016, 41, 313–318.

N. Maheshwari, S. Srivastava, and K.S. Rajan,

"Development of an Indoor Space Semantic Model and

Its Implementation as an IndoorGML Extension",

ISPRS International Journal of Geo-Information, vol. 8,

2019.

Ramón, A., Rodrıguez-Hidalgo, A. B., Navarro-Carrión, J.

T., & Zaragozı, B. (2013). A methodology for

evacuation route planning inside buildings using

geospatial technology. In Data Management and

Security. Applications in Medicine, Science, and

Engineering (Vol. 45, pp. 155-166). WIT Press.

Botsis, D., & Panagiotopoulos, E. (2021). Determination of

the shortest path in the university campus of Serres

using the Dijkstra and Bellman‐Ford algorithms.

ΧΩΡΟ-ΓΡΑΦΙΕΣ, 6(1).

Clementini, E., & Pagliaro, A. (2020). The Construction of

a Network for Indoor Navigation. In GISTAM (pp.

254-261).

Mirahadi, F., & McCabe, B. (2019). A real-time path-

planning model for building evacuations. In ISARC.

Proceedings of the International Symposium on

Automation and Robotics in Construction (Vol. 36, pp.

998-1004). IAARC Publications.

Liu, C., Mao, Z. L., & Fu, Z. M. (2016). Emergency

evacuation model and algorithm in the building with

several exits. Procedia Engineering, 135, 12-18.

Kolbe, T., Gröger, G., and Plümer, L. (2005). CityGML:

Interoperable access to 3D city models. In van Oost-

erom, P., Zlatanova, S., and Fendel, E., editors, Geo-

information for Disaster Management, pages 883–898,

Berlin, Heidelberg. Springer

Yang, L. and Worboys, M. (2015). Generation of

navigation graphs for indoor space. International

Journal of Geographical Information Science,

29(10):210–218

Modeling Simulating the Evacuation of a Building Based on Building Floor Plan and Evacuation Strategies

235

Mortari, F., Zlatanova, S., Liu, L., & Clementini, E. (2014).

" Improved Geometric Network Model"(IGNM): A

Novel Approach for Deriving Connectivity Graphs for

Indoor Navigation. ISPRS Annals of Photogrammetry,

Remote Sensing & Spatial Information Sciences, 2(4).

Sun, J., & Li, X. (2011, June). Indoor evacuation routes

planning with a grid graph-based model. In 2011 19th

International Conference on Geoinformatics (pp. 1-4).

IEEE.

Fu, M., & Liu, R. (2019). Automatic generation of path

networks for evacuation using building information

modeling. In Computing in Civil Engineering 2019:

Visualization, Information Modeling, and Simulation

(pp. 320-327). Reston, VA: American Society of Civil

Engineers.

Taneja, S., Akinci, B., Garrett, J. H., Soibelman, L., & East,

B. (2011). Transforming IFC-based building layout

information into a geometric topology network for

indoor navigation assistance. In Computing in Civil

Engineering (2011) (pp. 315-322).

Yenumula, K., Kolmer, C., Pan, J., & Su, X. (2015). BIM-

controlled signage system for building evacuation.

Procedia Engineering, 118, 284-289.

Liu, L., & Zlatanova, S. (2011). A" door-to-door" path-

finding approach for indoor navigation. Proceedings

Gi4DM 2011: GeoInformation for Disaster

Management, Antalya, Turkey, 3-8 May 2011.

Ma, J., Jia, W., & Zhang, J. (2017, May). Research of

building evacuation path to guide based on BIM. In

2017 29th Chinese Control And Decision Conference

(CCDC) (pp. 1814-1818). IEEE.

Xiong, Q., Zhu, Q., Zlatanova, S., Du, Z., Zhang, Y., &

Zeng, L. Y. (2015). Multi-level Indoor Path Planning

Method. International Archives of the

Photogrammetry, Remote Sensing & Spatial

Information Sciences.

GISTAM 2023 - 9th International Conference on Geographical Information Systems Theory, Applications and Management

236