Fuzzy Logic Framework for Local Accessibility Assessment based on

Built Environment Characteristics

Fatima Hanani

, Aziz Soulhi

Laboratory LASTIMI, CEDOC EMI, Mohamed V University, Rabat, Morocco

National Higher School of Mines, Rabat, Morocco

Keywords: Built environment, Fuzzy Logic, Land use, Local Accessibility.

Abstract: Researchers and decision-makers are increasingly interested in assessing the impacts of urban design and

transportation planning on local accessibility. The used accessibility measures present several issues and

limitations, namely: lack of understanding of accessibility concepts and technical and computational

complexity. In this paper, we present a new method to measure local accessibility. In this method, we use the

fuzzy logic approach. Our proposed method will measure local accessibility according to the three urban

characteristics, i.e., activity density, land use mix, and street design. This work has confirmed that accessibility

is an issue of urban design. In particular, it has shown that the combination of two urban characteristics,

namely activity density, and land use mix, is very determinant for accessibility. This work can serve as a

helpful tool for policymakers to understand and capture the interactions between accessibility, land use, and

travel behaviour.

1 INTRODUCTION

The global urban transition that has been underway

for several decades is phenomenal. It has put cities

and governments in front of unprecedented

challenges to provide urban infrastructure and

services, such as education, energy, transport and

water. In addition, climate change, environmental

constraints and resource scarcity have added more

stresses to cities.

Transportation is one of the most essential

services as it connects the different city areas and

allows people to access opportunities. According to

(The Global Mobility Report, 2017), the proposed

principles for sustainable transport have four goals,

efficiency, green mobility, safety, and universal

access. These objectives are also associated with

land-use planning, and their successful

implementation depends on the integration of

transport and urban planning. In this perspective, the

concept of accessibility is supposed to provide a basic

framework for this integration (Saghapour et al.,

2018). In (Zhang et al., 2015), accessibility is defined

by the spatial distribution of potential destinations,

the ease of reaching each one, and the extent, quality,

and character of the activities found there. Recently,

accessibility has gained ground in city institutions

that can use it most effectively as a planning tool

(Páez et al. 2012) and also as a tool to evaluate

(Saghapour et al., 2018) the effectiveness of policies

for land-use and transport planning.

To transform the concept of accessibility into a

measure used by decision-makers, an extensive

literature on accessibility measures exists. According

to (Miller, 2020), there are three categories of

accessibility measures: cumulative opportunities

measures, gravity-based measures, and utility-based

measures. These methods have several limitations.

Furthermore, despite the extensive literature on the

impact of the built environment on travel behavior,

there was relatively little evidence on the relationship

between accessibility and the built environment.

Therefore, we believe that writing accessibility in

terms of the characteristics of the built environment

shows the importance of integrating land use and

transport.

For this purpose, we will propose, in this work, a

new method based on fuzzy logic that allows to assess

the local accessibility (at street level) according to the

surrounding urban characteristics. As described in

(Ewing et al., 2010), the built environment has five

attributes, namely: density, diversity, design,

destination accessibility, and distance to

transportation. In this paper, we chose to study the

Hanani, F. and Soulhi, A.

Fuzzy Logic Framework for Local Accessibility Assessment based on Built Environment Characteristics.

DOI: 10.5220/0010730800003101

In Proceedings of the 2nd International Conference on Big Data, Modelling and Machine Learning (BML 2021), pages 187-195

ISBN: 978-989-758-559-3

187

following characteristics: density, diversity, and

design. This paper will answer the following

questions: Can sustainable accessibility (active

transport) be achieved through urban design? To what

extent can this opportunity serve city managers to

fulfill sustainable transport requirements?

This paper will be organized as follows. First of

all, we introduce the concept of accessibility and the

different methods used to measure it. After discussing

the limitations of existing accessibility measurement

methods, we will present our new measurement

method based on the fuzzy logic approach. Then, we

give a brief review of the literature (related works)

covering the impact of the built environment on

accessibility and travel behaviour. Besides, we

discuss the result of our method. Finally, we conclude

our paper by citing some perspectives for this work.

2 ACCESSIBILITY MEASURES

Accessibility has been the subject of much work

among researchers and actors (planners, decision-

makers, transportation, development). Despite

several years of active discussions, this concern is still

more present in the debate on spatial planning and

transport planning issues.

2.1 Local Accessibility Definition

Accessibility is a significant feature of urban areas

and often represents transport and land-use

objectives. Several scientific fields such as transport

planning, urban planning, and geography use this

concept, which plays an essential role in

policymaking (Karst et al., 2004). It can be a practical

tool for planning and evaluating transport and land

use planning (Saghapour et al., 2018).

There are several definitions of accessibility. We

quote some of them in the following. Firstly, in (Páez

et al. 2012), the author defines accessibility as the

possibility to reach opportunities (desired services

and activities) distributed in space and time.

Secondly, in (Zhang et al., 2015), the author describes

accessibility by the spatial distribution of potential

destinations, the ease of access (cost and time

savings, variety of transportation modes) to each

destination, and the extent (quality, diversity, and

character) of activities. Finally, according to (Karst et

al., 2004), accessibility is the extent to which land use

and transport systems enable (groups of) individuals

to reach activities or destinations using (a

combination of) transport mode(s).

In the light of the last definition, we can define local

accessibility as the extent to which land-use planning

allows (groups of) individuals to reach activities or

destinations utilizing active modes of transport

(walking and cycling). In this case, accessibility can

measure the impact of land use on the city’s

sustainability and individuals by offering them the

possibility to access activities by walking or cycling.

2.2 Existing Accessibility Measures

Accessibility measures generally consist of two

essential elements (Páez et al. 2012): the traveller'

cost (determined by the spatial distribution of

travellers and opportunities) and the quality/quantity

of opportunities. According to the accessibility

literature, there are three methods, which identify

three broad categories of indicators.

Cumulative opportunities: this measure counts the

number of opportunities reached within a given

access threshold (isochrone). This type of measure

focuses on the number of potential destinations or

opportunities rather than their distance and indicates

the choices available to residents (Karst et al., 2004).

Gravity-based measures: this measure relies on the

evidence that destinations become progressively less

attractive and less accessible as the cost (travel time,

effort, cost) increases. This phenomenon can be

considered by weighting each destination according

to a decay factor (gravity function) representing its

distance from the origin (McCahill et al., 2015).

Utility-based measures: this method refers to the

random utility theory. According to this, the

probability that an individual will make a particular

choice (e.g., destination, mode of transport) depends

on the utility of that choice relative to the utility of all

others (Zondag et al., 2015). This measure

corresponds to the log-sum of discrete choice models

applied to destination choice analysis (Páez et al.

2012).

2.3 Built-environment-based Measure

According to (Miller, 2020), the accessibility

measurement methods mentioned above present

several issues and limitations, namely: lack of

understanding of accessibility concepts (among

politicians, the public and non-modellers), technical

complexity, computational complexity, and lack of

standardized software availability and data.

Imprecision is another limitation of the different

methods of measuring accessibility. Thus, in the case

of arbitrary selection of the isochron, the imprecision

concerns the absence of differentiation between the

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

188

possibilities adjacent to the origin and those just

inside or outside the isochron. For the other methods,

as mentioned in (Handy et al. 1997), the parameters

of the impedance function have to be selected or

estimated. However, these parameters, which reflect

the relative importance of travel impedance in the

choice of destination, are based on aggregate travel

patterns rather than individual travel decisions.

Indeed, the individual in their evaluation of

accessibility uses quantitative and qualitative

information, making the underlying travel cost

(impedance) different from one individual to another

(Páez et al. 2012). In addition, as stated in (Zondag et

al., 2015), accessibility is seen as the main effect of

the transport system. Therefore, all methods

presented above calculate accessibility as a function

of transport system parameters (cost, travel time,

distance).

This work aims to highlight the fact that

accessibility is also an outcome of urban design.

Therefore, we present a new method, based on fuzzy

logic, to assess the accessibility according to the

urban characteristics of the area. We call it built-

environment-based accessibly measure. The use of

fuzzy logic seems relevant to us, given the limitations

mentioned above. Indeed, it is difficult to set up a

measure of accessibility with precise variables and

intervals. Furthermore, city managers need to read,

understand, and modify the rules of the accessibility

calculation model easily. Moreover, with fuzzy logic,

the accessibility measure is easy to understand and

interpret thanks to linguistic variables and human

reasoning.

3 BUILT-ENVIRONMENT AND

ACCESSIBILITY: RELATED

WORK

The substantial increase of the urban population,

urban sprawl and the distance from activities have

created a great need for travel that cannot be satisfied

by existing infrastructures and even by the

construction of others. This situation has led to the

high use of the private car. Consequently, cities face

recurrent congestion, pollution, social inequality,

road accidents, and increasing consumption of

gasoline (Toward Sustainable Mobility, 2019) (The

Global Mobility Report, 2017). As a result,

contemporary transportation focuses on changing

travel behaviour to reduce car travel and encourage

alternative modes, such as public transport, walking,

and cycling (Saghapour et al., 2018). In recent

decades, many studies have investigated and analysed

the interactions between urban form and traveller

behaviour. These studies have shown that travel

behaviour is impacted by socio-economic

characteristics (of the household) and built-

environment characteristics (of the surrounding area).

The latter is represented by the so-called 5D (density,

diversity, design, destination accessibility and

distance to transport) variables (Ewing et al., 2010).

In fact, the built environment impacts travel

behaviour through the degree of ease, the possibilities

offered to reach destinations, and the quality of

opportunities made available and accessible.

Therefore, accessibility, as defined above, is at the

heart of any change in travel behaviour.

Depending on the context, different studies on the

relationships between land use (5D variables) and

travel behaviour have focused on different transport

and travel parameters (trip frequency, distance

travelled, travel mode choices or total vehicle

kilometres travelled). However, we chose to limit our

research to these three characteristics (density,

diversity, design) considered by the scientific

literature to be the key factors that most influence

active transportation modes (at the local level) (Oakes

et al., 2007).

3.1 The Density of Activities

Density refers to the number of people, housing units,

jobs or floor area per unit area (Ewing et al., 2016). A

high density (residential, employment, other

activities, service, and leisure facilities) in a city will

reduce travel distances between residences,

workplaces and service facilities (Choi et al., 2020)

(Saghapour et al., 2016) on the one hand. On the other

hand, the complementary grouping of different

activities will help to better link different travel

objectives (Xia et al., 2020). Consequently, it will

limit energy consumption and vehicle emissions

(Yang et al., 2017) by creating walkable

environments and promoting public transport (Naess,

2012). As a result, residents of dense cities, with a

higher proportion of destinations within good

walking or cycling distance, can be expected to make

shorter daily trips on average than their counterparts

in less dense cities (Stevens, 2017). Therefore, this

can generate independence aims at the use of the

private car (Newman et al., 2006).

However, as discussed in (Deepty et al., 2019),

population or job density or even the aggregate

provide only a partial understanding and do not fully

capture the impact of the density of the set of

available activities on travel behaviour. Therefore, it

Fuzzy Logic Framework for Local Accessibility Assessment based on Built Environment Characteristics

189

would be wise to consider using a more

comprehensive variable to provide information on the

density of all activities in the area.

3.2 The Land Use Mix

Land use mix refers to the degree of concentration of

workplaces, shops, public administrations, cultural

events and recreational facilities (Song et al 2013). As

summarised by (Manaugh et al., 2013), since with a

single use of space, occupants will be obliged to use

motorized modes to get to their destinations, mixed-

use with complementarity will do the opposite. It will

allow the residents to walk or cycle to their

destination. According to several studies, the land-use

mix has several benefits to transportation, health,

economics and the environment (Manaugh et al.,

2013) (Hirt, 2016).

There are several methods for measuring land-use

mix (Song et al 2013). They all implicitly or explicitly

contain two concepts: distance and quantity. The

author in (Song et al 2013) surveyed the different

methods and classified them into two categories:

'Integral' and 'Divisional'. The first category of

measures, generally applied to small areas, tends to

reflect the balance of land use. However, the second

category, often applied to large geographic areas,

tends to reflect uniformity of land use.

3.3 Street Design and Network

Connectivity

As argued in (Brown et al., 2007), local urban design

principles, such as street configuration, availability of

sidewalks and bike lanes, and neighbourhood

aesthetic qualities, can influence the attractiveness of

non-motorised travel modes. In addition, the author

of (Ozbil et al., 2011) found that street network layout

is the leading independent variable affecting

pedestrian flow on streets. Furthermore, he argued

that shorter distances between intersections, smaller

block sizes and more direct paths encourage walking

and cycling. Indeed, according to (Ozbil et al., 2011),

the configuration of streets (connectivity) is

considered vital because it affects both the directness

of travel (making travel more or less efficient) and the

number of alternative routes, which has implications

for interest and safety. In other words, better network

connectivity can reduce travel distances for all

modes, including walking and cycling, and it can

provide more choices of routes. It should be

mentioned that there is essential literature dealing

with the measurement of connectivity. In (Dill et al.,

2004), the author evaluated several methods of

measuring connectivity (Block length, Block size,

Block density, Intersection density, Street density,

Connected Intersection Ratio, Percent four-way

intersections and Link-Node Ratio). In (Frank et al.,

2005), the author used the intersection density as a

measure of connectivity in his study at Atlanta. He

considered that areas with more than 30 intersections

per square kilometre are more walkable than other

areas. The author in (Litman, 2021) used Link-Node

Ratio as a connectivity indicator suggests that a link-

node ratio of 1.4 may be a good target for network

planning. Others found that values between 1.2 and

1.4 are good targets (Dill et al., 2004).

We have noted that the most used measures in the

literature are intersection density, street density,

connected node ratio and per cent four-way

intersection. In addition, the author in (Dill et al.,

2004) found a strong correlation between the first

three measures and suggested that Pedestrian Route

Directness (PRD) is the best measure to evaluate the

potential to encourage walking and cycling.

4 METHODOLOGY OF

LOCAL-ACCESSIBILITY

MEASUREMENT

4.1 Presentation of Fuzzy Logic

Fuzzy logic proposes a mathematical environment

built on the theory of fuzzy sets introduced in 1965

by Professor Lotfi A. Zadeh (University of California,

Berkeley). This approach attempts to simulate human

reasoning and allows the integration of imperfect data

in a decision process. As explained in (Hanani et al.,

2021), and described in figure 1, the basic

characteristics of fuzzy logic are the linguistic

variables, the universe of discourse, the function of

membership, and the fuzzy subset.

Figure 1: Fuzzy logic characteristics (Hanani et al., 2021)

Moreover, as presented in figure 2, the operating

principle of a fuzzy logic system includes three

Linguisticvariables

Thedescriptionof

thevariabletobe

studiedbyfuzzy

qualifierssuchas

(low,medium,high)

or(small,medium,

large).

Theuniverseof

discourse

Thephysicaldomain

associatedwiththe

variableunder

consideration

Thefunctionof

membership

Thisisthefunction

whichassociates,to

eachelementxofthe

universeofdiscourse,

thedegreetowhich

itbelongs(between0

and1)toafuzzy

subset.

Fuzzysubset

Thisisdefinedbytwo

things,auniverseof

discourseanda

functionof

membership..It

correspondingto

linguisticvalues.

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

190

phases. Namely: fuzzification, fuzzy inference

engine, and defuzzification.

Figure 2: fuzzy logic phases (Hanani et al., 2021)

4.2 Accessibility Assessment by Fuzzy

Logic

We suggest a new method for measuring local

accessibility that allows a significant level of spatial

disaggregation since we are interested in measuring

accessibility at the street level. Indeed, we will

describe local accessibility (output) as a function of

the three urban characteristics described above

(inputs), i. e., activity density, land-use mix and street

design. For this purpose, we follow the steps

illustrated in figure 3.

a) Fuzzification: we start by defining fuzzy

subsets and membership functions for each variable

of our fuzzy system (Input and Output). Then we

translate the different variables into fuzzy language.

b) Fuzzy inference: this is where we apply

human reasoning. This phase consists of two steps.

The first step is to build decision rules and find the

membership rule of the conclusion for each of them.

The second step consists of the aggregation of the

conclusions. For this phase, we use the Mamdani

inference mechanism.

c) Defuzzification: this final phase extracts a

real value from the fuzzy subset resulting from the

previous step. We chose to use the centre of gravity

method because it considers the entire final

membership function when calculating the final

result.

4.2.1 Definition of Linguistic Variables

To model our system to evaluate local accessibility,

we have defined the four variables. For each variable,

we determine

the fuzzy subset and the membership

functions.

a) Activity density index

As we pointed out above, it would be more relevant

to think about using a more exhaustive variable to

evaluate the influence of activity density on local

accessibility. In this regard, we have chosen to assess

the density exhaustively, considering all the existing

activities in the studied area (residential, jobs, other

activities, services, leisure facilities). Theoretically,

we cannot define the optimal distribution of each type

of activity that would lead to the ideal density (of

activities) in an area. Therefore, we will introduce a

reference area where the activity density is optimal.

Inspired by (Song et al 2013), and to have a

normalized variable, we have introduced a new

measure that will inform us about the degree of

dispersion of our study area compared to the

reference area. We call it the activity density index

(ADI). We assume that R is the reference area and k

is the number of activity types present in this area. For

each activity type i (from 1 to k), the density

percentage of each activity type i is 𝑟



with

∑

𝑟



1.

For a zone X with density percentages of each

type of activity 𝑥



(

∑

𝑥



1), we determine ADI by

measuring the dispersion of the density of X to R (the

average deviation from the reference). The ADI tells

us how much the activity density of area X deviates

from our reference area. We can calculate the activity

density index (ADI) of area X as follows:

𝐴𝐷𝐼





1 𝑟



𝑥



𝑟



1





In case of a wide deviation from the reference, the

value

𝑥



𝑟



is close to 1. Therefore, the activity

density index is close to 0. Assuming a density close

to the reference, the value

𝑥



𝑟



is close to 0.

Therefore, the activity density index is close to 1. For

this variable, we choose the following subset:

 High activity density: when 𝐴𝐷𝐼 is high

than 0,8

 Medium activity density: when the 𝐴𝐷𝐼 is

amount 0,6

 Low activity density: when the 𝐴𝐷𝐼 is less

than 0,4

b) Land Use Mix

Based on (Song et al 2013), we choose to use the

entropy index, which is the most used measure to

evaluate the land use mix. Its formula is as follows:

𝐿𝑈𝑀  𝑃







ln𝑃



 ln



𝑘



 2

Where 𝑃



the percentage of each land-use type j in the

area, and 𝑘 is the number of land-use types

(categories of interest). The Entropy Index varies

from 0 (least mixed area) to 1 (most mixed area)

FUZZiFICATION

Transformingtheactual

variablestobestudied

(inputandoutput)into

linguisticvariablesby

assigningthemdegreesof

membershiptofuzzy

subsets

INFERENCE ENGINE

Consistofthe

inferenceruleswhich

isbuiltonexpert

knowledgebases

Allowgenerationof

thefuzzyoutputs

fromthefuzzyinputs

DEFUZZIFICATION

Transformthefuzzy

setresultofinference

enginetosinglenet

resultthatrepresents

theoutputofthe

fuzzysystem

Fuzzy Logic Framework for Local Accessibility Assessment based on Built Environment Characteristics

191

(Litman, 2021). According to (Litman, 2020), we

choose the following subset:

 High Land use mix: when 𝐿𝑈𝑀 is high than

0,7

 Medium Land use mix: when the 𝐿𝑈𝑀 is

between 0,5 and 0,7

 Low Land use mix: when the 𝐿𝑈𝑀 is around

0,3

c) Street Design and Connectivity

For our study, since our objective is to see how the

built environment can improve local accessibility and

encourage active modes of transportation, we chose

to use Pedestrian Route Directness (PRD) to measure

street connectivity. As said above, it may be a better

measure that can inform the promotion of cycling and

walking than other measures. The PRD is obtained by

the ratio between the shortest Route distance (𝐷



) and

the Straight-line distance (𝐷



) for two selected points.

𝑃𝑅𝐷

𝐷



𝐷



3

The lowest possible value is 1, where the shortest

Route distance ( 𝐷



) is the same distance as the

Straight-line distance (𝐷



). Values further than one

(1) are not recommended because it indicates that the

route is not direct and there are several changes of

direction to reach the destination.

Based on (Dill et al., 2004), we define our subset for

connectivity index as follows:

 High Street design & network connectivity:

when 𝑃𝑅𝐷 is less than 1,5

 Medium Street design & network

connectivity: when the 𝑃𝑅𝐷 is between 1,5

and 1,8

 Low Street design & network connectivity:

when the 𝑃𝑅𝐷 is higher than 1,8

d) Local Accessibility

We have noticed that there is no standard for

assessing local accessibility. Indeed, we can state

whether one area is more accessible than another by

comparing accessibility (regardless of the method

used for the calculation). But we certainly cannot

determine the perfect accessibility level for an area.

Therefore, we will introduce a reference area where

accessibility is optimal. We use the accessibility

value of this area as a baseline to assess the

accessibility of any studied zone.

We will assume that the accessibility of the

reference area is 𝐴



, and the accessibility of the

studied area is 𝐴



. We define a local accessibility

index (LAI) as the ratio between 𝐴



and 𝐴



𝐿𝐴𝐼

𝐴



𝐴



4

The highest possible value of LAI is 1, where the

accessibility of the studied area (𝐴



) is the same as

the reference zone (𝐴



). Small values than one (1) are

not recommended because it indicates that the studied

area is not well accessible. We define our subset for

local accessibility index as follows:

 High accessibility when 𝐿𝐴𝐼 is high than 0,7

 Medium accessibility when 𝐿𝐴𝐼 is between 0,5

and 0,7

 Low accessibility when 𝐿𝐴𝐼 is less than 0,5

4.2.2 Membership Functions and Inference

Rules

Figure 3: Membership Functions.

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

192

Figure 4: Example of Inference Rules.

Through the above section, we have built our fuzzy

system that measures accessibility in regards to three

variables, i. e. activity density, land-use mix and

street design. We define the universe of discourse, the

fuzzy subset (High, Medium and Low) and the

membership functions for each variable.

Figures 3 and 4 give an overview of our fuzzy

system for built-environment accessibility

measurement.

5 DISCUSSION

5.1 Result Interpretation

After the previous steps, our fuzzy system for

evaluating the local accessibility is now ready. We

proceed to the analysis and interpretation of the

defuzzification results. To do that, we will analyse the

surface graphs for three cases.

5.1.1 Case 1: Fixed ADI in an Average

Value

In this case, we set the density to an average value

(0.6), and we see how our system reacts.

We can see in Figure 5 that when the density is

medium, we can only expect a medium value of

accessibility. Therefore, when the PRD is low,

accessibility remains low whatever the value of the

LUM variable. This result shows that it is crucial to

consider ease of access when studying and plaining

local accessibility.

5.1.2 Case 2: Fixed LUM in an Average

Value

As per figure 6, when we set the LUM to a medium

value (0,5), we notice that this case presents the same

result as the previous one. The maximum

accessibility value we can expect is medium.

Therefore, when the PRD is low, accessibility

remains low whatever the value of the ADI variable,

which confirms the link between accessibility and the

ease of reaching a destination.

According to case 1 and case 2, we conclude that

the land-use mix has the same impact as the activity

density on local accessibility.

Figure 5: Surface View for the case N°1.

Figure 6: Surface View for the case N°2

5.1.3 Case 3: PRD Fixed in an Average

Value

As we can see in figure 7, in contrast to the two

previous cases, the most important remark is that

accessibility can reach high values but with one

crucial condition: both ADI and LUM must be

increased. In addition, this graph shows two main

findings:

- Local accessibility can be medium when at least one

of the two has a medium value;

- Local accessibility is low when at least one of the

two has a low value.

This result is consistent because when the density is

low, there is no reason to talk about the land-use mix.

Moreover, when the density is high, and the land use

Fuzzy Logic Framework for Local Accessibility Assessment based on Built Environment Characteristics

193

mix is down, we conclude that the area is mono-

activity. In both cases, the accessibility remains low,

except when PRD become high. In this case,

accessibility takes a medium value (figure 8).

Furthermore, we can conclude that a combination of

ADI and LUM characteristics are crucial and relevant

to achieve high local accessibility.

Figure 7: Surface View for the case N°3

Figure 8: Surface View for PRD high value

5.2 Contributions and Limitations

Although the commonly used method for measuring

local accessibility is the cumulative opportunity

method. In this paper, we chose to extend the scope

by considering the surrounding land use environment.

We decided to evaluate local accessibility using the

principles of the gravity method, incorporating an

impedance function capturing the access conditions

to each opportunity (PRD). This method allowed us

to consider the two main components of accessibility

(Karst et al., 2004): the ease of walking (or cycling)

to reach a destination and the quantity and spatial

distribution of opportunities. The first component is

represented, in our model, by the characteristic ‘Street

Design and Connectivity’ (PRD), and the second by

the combination of the two other urban

characteristics, i.e., Activity Density Index and the

Mixed-Use Index. We have suggested a model that

allows a high level of disaggregation to capture small-

scale design features (street scale), also evaluate non-

motorized trips. Therefore, our model can help city

decision-makers predict the full impacts of land use

management strategies on improving local

accessibility. Namely, densification of activities,

mixing activities and bringing them closer together by

improving walking and cycling conditions and

pedestrian-friendly environments.

To be relevant and practical, a model must be

exhaustive and consider all the factors and elements

that can influence accessibility. However, although

our model considers the two main components of

accessibility (opportunity and ease of access), it does

not consider how these two components are perceived

and used by individuals with different characteristics

(Páez et al. 2012). Therefore, our model deals only

with local accessibility, and it does not consider

regional accessibility. It only concerns active modes

of transport and not others.

6 CONCLUSIONS

It is widely recognised that accessibility is one of the

main effects of the transport system (Zondag et al.,

2015). However, this paper shows that accessibility is

also a matter of urban planning. In particular, we have

found that local accessibility can be defined by three

urban characteristics, i.e., activity’s density, land-use

mix and street design.

This paper presents a conceptual framework and a

new method to measure local accessibility. Based on

one of the tools of artificial intelligence, which is

fuzzy logic. This method is easy to understand and

interpret thanks to the use of linguistic variables and

human reasoning. It also showed that accessibility is

more affected by the two main characteristics, namely

activity density and land-use mix.

Furthermore, through this work, our objective is

to participate in the collective effort of researchers to

propose a model that allows transportation and land

use planners in cities to predict how their policies and

decisions can improve local (active) accessibility and

sustainable development.

In the perspectives, and to complete this work, we

intend to extend our model to treat the question of

accessibility globally and to consider the regional

dimension. Also, our work can be improved by

testing it with actual data to calibrate it.

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

194

REFERENCES

Brown, B.B., Werner, C.M., Amburgey, J.W. and Szalay,

C., 2007. Walkable route perceptions and physical

features: Converging evidence for en route walking

experiences Environment and Behavior Vol 39(1),

pages 34–61.

Choi, S., Choo, S. and Kim, S., 2020. Exploring the

influences of compact development on zone-based

travel patterns: a case study of the Seoul metropolitan

area, Transportation Letters, 12:5, 316-328,

Chris McCahill, Mary Ebeling, 2015. Tools for measuring

accessibility in an equity framework, Congress for the

New Urbanism 23rd Annual Meeting Dallas, Texas

April 29 – May 2.

Deepty. J, Geetam. T., 2019. Measuring density and

diversity to model travel behaviour in Indian context,

Land Use Policy, Volume 88, 104-199.

Dill, J., 2004, Measuring Network Connectivity for

Bicycling and Walking, TRB Annual Meeting.

Ewing, R. and Cervero, R., 2010. Meta Analyse: Travel and

the built environment, Journal of the American

Planning Association, Vol. 76, No. 3, pp. 265-294

Ewing. R., Hamidi. S. and Grace. JB., 2016. Compact

Development and VMT—Environmental Determinism,

Self-Selection, or Some of Both? Environment and

Planning B: Planning and Design, Vol 43, pp. 737-755

Frank, L.D., Schmid, T.L., Sallis, J.F., Chapman, J.,

Saelens, B.E., 2005. Linking objectively measured

physical activity with objectively measured urban form:

Findings from SMARTRAQ, American Journal of

Preventive Medicine, 28 (2 SUPPL. 2), pp. 117-125

Global Mobility Report, 2017.

https://sustainabledevelopment.un.org/content/docume

nts/2643Global_Mobility_Report_2017.pdf,

15/04/2021 22h30

Hanani, F., Soulhi, A., 2021 improving traffic congestion

assessment by using fuzzy logic approach. Journal of

Theoretical and Applied Information Technology,

Vol.99. No 3 pp 625-638.

Handy, S L, Niemeier, D A, 1997. Measuring accessibility_

an exploration of issues and alternatives, Environment

and Planning A, volume 29, pages 1175-1194.

Hirt, S. A., 2016. Rooting out mixed use: Revisiting the

original rationales, Land Use Policy, Volume 50, Pages

134-147,

Karst T. Geurs, Bert van Wee, 2004. Accessibility

evaluation of land-use and transport strategies: review

and research directions. Journal of Transport

Geography, Volume 12, Issue 2, Pages 127-140,

Litman, T., 2020. Evaluating Accessibility for Transport

Planning. https://www.vtpi.org/access.pdf, 23/04/2021,

13h30

Litman, T., 2021. Land Use Impacts on Transport,

https://www.vtpi.org/landtravel.pdf, 26/04/2021

15h30.

Manaugh, K., & Kreider, T., 2013. What is mixed use?

Presenting an interaction method for measuring land

use mix. Journal of Transport and Land Use, 6(1), 63-

72.

Miller, E., 2020, “Measuring Accessibility: Methods and

Issues”, International Transport Forum Discussion

Papers, No. 2020/25, OECD Publishing, Paris.

Naess, P., 2012. Urban form and travel behaviour:

experience from a Nordic context. Journal of Transport

and Land Use, 5(2)

Newman. P. W. G. and Kenworthy. J. R., 2006. Urban

Design to Reduce Automobile Dependence”, Opolis:

An International Journal of Suburban and Metropolitan

Studies, Vol. 2, No. 1, 2006, pp. 35-52.

Oakes JM, Forsyth A, Schmitz KH., 2007. The effects of

neighborhood density and street connectivity on

walking behavior: The Twin Cities walking

study. Epidemiol Perspect Innov.;4:16.

Ozbil. A., Peponis. J. and Stone. B., 2011. Understanding

the link between street connectivity, land use and

pedestrian flows, urban design International, vol 16},

pp 125-141

Páez, A, Scott, D. M., Morency, C, 2012. Measuring

accessibility: positive and normative implementations

of various accessibility indicators, Journal of Transport

Geography,Volume 25 , Pages 141-153,

Saghapour, T, Moridpour, S. & Thompson, R, 2016. Public

transport accessibility in metropolitan areas: A new

approach incorporating population density, Journal of

Transport Geography, Volume 54, Pages 273-285,

Saghapour, T, Moridpour, S. & Thompson, R, 2018.

Sustainable transport in neighbourhoods: effect of

accessibility on walking and bicycling.

Transportmetrica A: Transport Science

Song, Y., Merlin, L., Rodriguez, D., 2013. Comparing

measures of urban land use mix, Computers,

Environment and Urban Systems, Volume 42, Pages 1-

13,

Stevens, M. R., 2017. Does Compact Development Make

People Drive Less? Journal of the American Planning

Association, 83:1, 7-18,

Toward Sustainable Mobility, 2019.

https://www.sum4all.org/data/files/gra-

globalroadmapofaction-press.pdf, 10/04/2021, 12h20

Xia, C., Gar-On Yeh, A., Zhang, A. 2020. Analysing spatial

relationships between urban land use intensity and

urban vitality at street block level: A case study of five

Chinese megacities, Landscape and Urban Planning,

Volume 193, 103669,

Yang, B., Xu, T., Shi, L. 2017. Analysis on sustainable

urban development levels and trends in China's cities,

Journal of Cleaner Production, Volume 141, Pages 868-

880,

Zhang, C., Man, J., 2015. Examining Job Accessibility of

the Urban Poor by Urban Metro and Bus: A Case Study

of Beijing. Urban Rail Transit 1, 183–193.

Zondag, B., De Bok, M., Geurs, K. T., Molenwijk, E., 2015.

Accessibility modeling and evaluation: The TIGRIS

XL land-use and transport interaction model for the

Netherlands, Computers, Environment and Urban

Systems, Volume 49, Pages 115-125.

Fuzzy Logic Framework for Local Accessibility Assessment based on Built Environment Characteristics

195