IEEE - 61001

Optimizing Electric Vehicle Charging Station

Deployment: A Predictive Modeling and
Optimization Approach
2024 15th International Conference on Computing Communication and Networking Technologies (ICCCNT) | 979-8-3503-7024-9/24/$31.00 ©2024 IEEE | DOI: 10.1109/ICCCNT61001.2024.10725354

Mohd Sarfaraz Faiyaz Garima Srivastava Sachin Kumar

Amity University Lucknow Campus Amity University Lucknow Campus Amity University,Lucknow Campus
[email protected] [email protected] [email protected]

Abstract — Owing to extensive environmental advantages, term growth of the EV industry [3,4]. The quantity and
energy optimized processes, and amazing technological placement of charging stations are significant elements
innovations, Electric Vehicles (EVs) have become an important influencing the growth of the battery-powered mobility
domain for research and development. EVs utilize stored industry, linked to various economic and social issues.
electricity power from rechargeable batteries rather than
conventional ICE vehicles, which run on gasoline or diesel fuel Strategically placing charging stations can mitigate the
to power their engines. increasing adoption of electric vehicles primary barrier to mobility carbon reduction, which is the
(EVs) has placed a strain on the current infrastructure for annoyance created by infrastructure investment and charging
charging, leading to the need for efficient deployment of EV time [5]. This issue is of current research interest and has
charging stations. Work done is an effort to identify the areas enormous consequences for the EV revolution, including
with the greatest potential for installing EV charging points and reducing EV owners' range anxiety, optimizing EV charging
suggest the fastest and most affordable charging options. It point use, and reducing EV owners' trip and waiting time. As
evaluates current EV adoption trends and utilizes curve-fitting EVs spread across multiple locations, the problems associated
to predict future demand suited for EV charging facilities. It with this issue evolve, and the requirement for solutions at
then leverages the Gurobi optimization solver to address cost various geographies and sizes will grow dramatically in the
minimization by allocating demand to charging stations next decades. This study aims to optimize the deployment of
strategically. The results demonstrate the effectiveness of this
EV charging stations to meet the growing demand for EVs
approach in optimizing EV charging station deployment,
achieving an accuracy of over 93%. Notably, Scenario 1
efficiently.
achieved an optimal solution with a cost of INR 3,978,066/-, II. LITERATURE REVIEW
highlighting the potential cost savings this method offers.
Optimizing the deployment of EV charging stations is essential Vazifeh et al. in 2019 proposed a data-driven
to ensure a smooth shift to ecofriendly transportation and methodology for optimizing the deployment of Electric
satisfy the rising demand for it. Vehicle (EV) charging stations, leveraging pervasive
mobility data [6]. Their approach addresses the increasing
Keywords— Electric Vehicles (EVs), Infrastructure, demand for public charging infrastructure amidst
Charging Stations, Optimization, Sustainability
advancements in battery technology and the ubiquity of
I. INTRODUCTION location-tracking devices. By formulating the problem as a
discrete optimization task, they aim to minimize drivers'
The transportation industry is responsible for discomfort while efficiently covering the demand region.
approximately eighteen percent of the greenhouse gas Employing greedy and genetic algorithms, they demonstrate
emissions from the combustion of fuel, making it a significant significant reductions in both discomfort and required
contributor to global warming. Decarbonizing mobility charging stations compared to randomized solutions, with
requires a sizable market proportion of electric vehicles (EVs)
the genetic algorithm outperforming. Moreover, they assess
powered by clean energy [1,2]. With the increase in the
adoption of EVs, the demand for charging is also rising. the methodology's robustness over time, showing consistent
Currently, only 10% of drivers choose electric vehicles when performance in later months. The study underscores the
purchasing a new vehicle, but this number is expected to rise potential of data-driven strategies in optimally locating
quickly. To support the expansion of the EV industry, it is public charging facilities within urban areas, contributing to
crucial to accelerate technology and infrastructure the field of Intelligent Transportation Systems and EV
development. infrastructure development.
Another study by Campaña and Inga in 2023 presente a
The popularity of new energy electric cars has increased theoretical model for optimal deployment of fast-charging
significantly, making it essential to satisfy the charging stations for electric vehicles (EVs), considering factors such
requirements of electric vehicles efficiently. As a result,
as sizing the electrical distribution network and traffic
determining the ideal allocation of chargers for electric
conditions [7]. Their model leverages data from
vehicles has become an important research subject. This study
examines the issues influencing charging station distribution OpenStreetMap and is validated using CymDist software. By
from various perspectives, such as environmental advantages, predicting EV numbers and minimizing vehicular flow
and proposes an optimum supply chain and algorithm for trajectory through a traffic rerouting system, the model
electric car charging stations. The allocation of charging addresses challenges in ensuring interoperability and
infrastructure plays a significant role in driving the EV reducing waiting times at charging stations. It incorporates
market's development. It is critical to define the ideal stochastic analysis of vehicular flow, capacitated
concentration of EV charging infrastructure according to multicommodity flow problem solution, analysis of optimal
geographic locations and market criteria to ensure the long- terminal numbers considering loading times, and validation

15th ICCCNT IEEE Conference,

Juneon
Authorized licensed use limited to: New York University. Downloaded 24-28, 2024, 07,2024 at 01:29:39 UTC from IEEE Xplore. Restrictions apply.
November
IIT - Mandi, Kamand, India
 IEEE - 61001

through CymDist. The study highlights the industrial determining optimal placement and additional aspects are
applicability of the model, particularly for charging station elaborated in subsequent sections.
infrastructure providers in developing countries, showcasing
its potential in optimizing EV charging infrastructure IV. ABOUT DATASETS
development within smart city frameworks. The datasets used are:
Another study by Gulbahar et al. in 2023 in their ork i. The dataset "Demand_History.csv" comprises a
addressed the crucial need for an optimal charging time-series record of EV charging demand across a
infrastructure for electric vehicles (EVs), focusing on region, delineated by a demand map featuring 64x64
intercity travel along highways [8]. While existing research equidistant points. It encompasses the index, x-
often concentrates on urban areas, this study proposes a coordinate, and y-coordinate for each demand point,
decision model to determine the optimal placement of along with the demand history spanning from 2010
charging stations on national highways in Türkiye. They to 2018.
utilize a novel approach to estimate the future number of
ii. "Existing_EV_infrastructure_2018.csv" contains
EVs, crucial for their facility location model. Through
particulars regarding the prevailing EV
optimization techniques, the model identifies suitable
infrastructure as of 2018. It encompasses the index,
locations for charging stations and determines the required x-coordinate, and y-coordinate for each of the 100
number of chargers to meet customer demand. By ensuring parking locations (termed supply points), alongside
EV drivers on highways encounter minimal charging the maximum number of parking slots available at
problems, the study contributes to the advancement of each location and the count of SCS and FCS units
sustainable transportation. This research underscores the already operational at these sites.
importance of extending charging infrastructure beyond
urban centers, facilitating the widespread adoption of EVs The objective is to use the provided data to forecast EV
and promoting environmentally friendly mobility solutions. charging demands and then optimize the EV infrastructure to
meet this forecasted demand while adhering to practical
III. PROBLEM STATEMENT constraints. All quantities/values presented in the
aforementioned datasets lack dimensionality. The demand
and supply data do not correspond to any actual location.
Nonetheless, the dataset exhibits characteristics typical of
practical use cases. Notations and constants used are:
i. 𝑖 = Index representing the demand point. 𝑖 ranges
from 0 to 4095.
ii. 𝑗 = Index representing the supply point (parking
location). 𝑗 ranges from 0 to 99.
iii. 𝐷𝑖 = ev charging demand at 𝑖 th demand point.
iv. 𝑆𝐶𝑆𝑗 = number of slow charging stations at 𝑗 th
supply point
v. 𝐹𝐶𝑆𝑗 = number of fast charging stations at 𝑗th supply
point
Fig. 1. Problem description - map is used only for describing the problem. vi. 𝑃𝑆𝑗 = total parking slots available at 𝑗th supply point
vii. 𝐶𝐴𝑃 SCS = charging capacity of a slow charging
The problem involves projecting Electric Vehicle (EV) station = 200
charging needs and optimizing EV infrastructure using a
viii. 𝐶𝐴𝑃 FCS = charging capacity of a fast-charging
geographic region divided into equal-sized blocks. Each
block contains specific total EV charging requirements, station = 400
termed demand points, illustrated as red circles in Fig. 1. ix. 𝑆𝑚𝑎𝑥 j = maximum supply that can be given from 𝑗th
These demand points collectively form a requirement map for supply point = (𝐶𝐴𝑃SCS * 𝑆𝐶𝑆𝑗)+ (𝐶𝐴𝑃FCS * 𝐹𝐶𝑆𝑗)
the region, with forecasting achievable for each demand point x. 𝐷𝑖𝑠𝑡 ij = distance between 𝑖 th demand point and 𝑗 th
utilizing historical demand maps. supply point (distance matrix)
xi. 𝐷𝑆ij = How much demand of the 𝑖 th demand point is
EV charging stations are typically situated at satisfied by the 𝑗 th supply point (Demand-Supply
predetermined public parking locations, denoted as supply matrix)
points, represented by green stars in Fig. 1. These locations
have a fixed number of parking slots, serving as potential sites A. Objective:
for EV charging station installation. Charging stations come The objective of the EV infrastructure optimization
in two types, slow charging station (SCS) and fast charging challenge is to minimize the overall cost, which comprises
station (FCS), with the maximum supply determined based three distinct components. Cost of Customer Dissatisfaction
on the number of SCS and FCS installed at a parking location (Cost CD): This component reflects the expense associated
and their respective charging capacities. The aggregation of with customer dissatisfaction, determined by the distance
supply points in a geographic region forms a supply map over customers must travel to fulfill their EV charging needs. It is
the area. By utilizing the demand and supply maps, along with calculated as the sum of the distances multiplied by the extent
demand-supply constraints and objectives, the optimal to which demand is satisfied at each point.
placement of EV charging stations can be determined. This
designed EV infrastructure is tailored to effectively meet the 𝐶𝑜𝑠𝑡𝐶𝐷 = ∑ 𝐷𝑖𝑠𝑡𝑖𝑗 ∗ 𝐷𝑆𝑖𝑗
projected demand. Further details on the methodology for 𝑖.𝑗

15th ICCCNT IEEE Conference,

Juneon
Authorized licensed use limited to: New York University. Downloaded 24-28, 2024, 07,2024 at 01:29:39 UTC from IEEE Xplore. Restrictions apply.
November
IIT - Mandi, Kamand, India
 IEEE - 61001

Cost of Demand Mismatch (Cost DM): This element accounts optimization problems in operations research, finance, and
for the cost resulting from discrepancies between forecasted scheduling [10].
and true demand values. It is computed as the sum of the
absolute differences between forecasted and actual demand One of the key advantages of MIP is its ability to handle
values for each demand point. discrete decisions that need to be made [11]. For example, in
supply chain management, a company may need to select the
optimal combination of suppliers and transportation routes to
𝐶𝑜𝑠𝑡𝐷𝑀 = ∑ 𝑎𝑏𝑠(𝐷𝑓𝑜𝑟𝑒𝑐𝑎𝑠𝑡,𝑖 ∗ 𝐷𝑡𝑟𝑢𝑒,𝑖 )
minimize costs while meeting demand. MIP models can also
𝑖
handle constraints that are difficult or impossible to express
Cost of Infrastructure (Cost IF): This component pertains to with continuous variables alone, such as minimum or
the operational, maintenance, and amortized capital costs maximum production capacities or minimum order
incurred in owning EV infrastructure. It is calculated as the quantities. To solve a MIP problem, a software program such
sum of the total number of slow charging stations (SCS) and as Gurobi can be used. These programs use mathematical
fast charging stations (FCS), with the cost of FCS being 1.5 algorithms to find the optimal solution to the problem, given
times that of SCS. the objective function and constraints. The solution will
specify the optimal values of the integer and continuous
𝐶𝑜𝑠𝑡𝐼𝐹 = ∑(𝑆𝐶𝑆𝑗 + 𝑟 ∗ 𝐹𝐶𝑆𝑗 ) variables, which can be used to make decisions in the real
𝑗 world. However, MIP can also be computationally intensive
and time-consuming, especially for large-scale optimization
Where, 𝑟=1.5 is the ratio of the cost of fast charging problems with many variables and constraints. Therefore, it
station to that of slow charging station. is important to carefully design the MIP model to ensure that
The constants used in the calculations are: a = 1, b = 25, c it is both accurate and efficient. This may involve simplifying
= 600. the model, reformulating the problem, or using specialized
algorithms or techniques to improve performance. With
𝐶𝑜𝑠𝑡 = 𝑎 ∗ 𝐶𝑜𝑠𝑡𝐶𝐷 + 𝑏 ∗ 𝐶𝑜𝑠𝑡𝐷𝑀 + 𝑐 ∗ 𝐶𝑜𝑠𝑡𝐼𝐹 careful design and appropriate use, MIP can be a powerful
B. Constraints: tool for solving complex optimization problems in a variety
of applications.
We have taken certain constraints for optimizing the EV
charging infrastructure. V. PROPOSED MODEL
i. All values of the demand-supply matrix (DSij) must A. Tasks to do:
be non-negative.
i. The initial task is to optimize the code by
ii. All values of the number of slow (SCSj) and fast eliminating any mistakes or unneeded
charging stations (FCSj) must be a non-negative components to improve its readability and
integer. efficiency.
iii. Sum of slow (SCSj) and fast charging stations ii. The second task is to construct pipelines and
(SCSj) must be less than or equal to the total parking classes for the code, which involves breaking the
slots (PSj) available at each jth supply point. code into different modules and encapsulating
each module's functionality into classes. This
iv. Expansion of EV infrastructure is limited to building approach improves code maintenance,
upon the existing 2018 infrastructure. Thus, the reusability, and scalability.
number of slow charging stations (SCS) and fast
charging stations (FCS) at each supply point (j) must B. The thought process:
either increase or remain unchanged annually. i. The first task is to use a curve fitting technique to
v. The aggregate fraction of demand satisfied by each predict the demands based on the existing trend. The
supply point (j) must not exceed the maximum objective function used is ax3 + bx2 + cx + d.
available supply at that point. ii. The second task involves solving a cost
minimization problem using Gurobi in Python to
∑ 𝐷𝑆𝑖𝑗 < 𝑆𝑚𝑎𝑥𝑗 allocate the demand from each demand point to EV
𝑖 charging stations optimally.
vi. (Sum of fractional) Forecasted demand at each ith
demand point must exactly be satisfied.

∑ 𝐷𝑆𝑖𝑗 = 𝐷𝑓𝑜𝑟𝑒𝑐𝑎𝑠𝑡,𝑖
𝑗

Our Approach: Mixed Integer Programming (MIP) using

Gurobi
Mixed Integer Programming (MIP) is a type of
mathematical optimization problem that combines both
continuous and integer variables [9]. In MIP problems, some
variables are restricted to taking on integer values only, while
others can take on any value within a specified range. This
makes MIP a versatile tool for solving a variety of Fig. 2. Curve fitting technique to predict the demands based on the
existing trend.

15th ICCCNT IEEE Conference,

Juneon
Authorized licensed use limited to: New York University. Downloaded 24-28, 2024, 07,2024 at 01:29:39 UTC from IEEE Xplore. Restrictions apply.
November
IIT - Mandi, Kamand, India
 IEEE - 61001

The statistical technique known as curve-fitting is Subject To:

utilized to determine the best curve that represents the 𝐷𝑆𝑖𝑗 ≥ 0, 𝐷𝑆𝑖𝑗 ∈ 𝑹
relationship between two or more variables. By fitting a
mathematical function to a set of observed data points, curve- 𝑆𝐶𝑆𝒋 ≥ 𝑆𝐶𝑆𝑝𝑟𝑒𝑣,𝑗 , 𝑆𝐶𝑆𝑗 ∈ 𝑰
fitting aims to identify the function that best represents the 𝐹𝐶𝑆𝒋 ≥ 𝐹𝐶𝑆𝑝𝑟𝑒𝑣,𝑗 , 𝐹𝐶𝑆𝑗 ∈ 𝑰
underlying trend or pattern in the data. In the case of EV
charging stations, existing demand at each location is used as
data points, and the curve-fitting process is employed to find 𝐷𝑖 = ∑ 𝐷𝑆𝑖𝑗
the curve that describes this relationship most accurately 𝑗
depicted in figure 2. The resulting curve can then be utilized 𝑆𝐶𝑆𝑗 + 𝐹𝐶𝑆𝑗 ≤ 𝑃𝑆𝑗
to forecast future demand for EV charging stations at each
location, which can assist in optimizing the deployment of ∑ 𝐷𝑆𝑖𝑗 ≤ 𝑆𝑚𝑎𝑥𝑗
charging stations to meet the rising demand for electric
𝑖
vehicles [12,13].
F. Comparision: Number and availability of the charging
Gurobi is a widely used optimization solver that can stations
solve various optimization problems, including mixed integer
programming (MIP) problems [14]. MIP problems are Presence of charging stations in total was also estimated
optimization problems where some variables are constrained emloying equation given below, the number of charging
to take on integer values, making them more challenging to stations per 1 person (NCSP),
solve than linear programming problems. Gurobi offers a NCSP = NCS/P
Python API that enables users to build and solve MIP models
Further the number of people per 1 charging station (NPCS),
within a Python environment. The Gurobi Python API
were calculated in this study,
provides an array of functions and classes that simplify
building and solving MIP models, making it easy for users NPCS = P/NCS
with limited optimization expertise to use. where:
NCSP: number of charging stations per 1 person;
C. The variables are set up as: NPCS: number of people per 1 charging station;
i. num_fcs : integer value representing the number of NCS: number of charging stations in specific country;
fast charging stations available at each supply point. P: population of specific country (number of citizens).
ii. num_scs : integer value representing the number of
slow charging stations available at each supply Table 1depicts the minimum number of individuals per
point. single charging station. [15]

iii. Allocation Matrix: a floating-point value indicating Table 1. Top countries with the minimum number of individuals per single
the amount of demand from demand point i that can charging station. [15]
be supplied by supply point j.
D. The constraints of the problem:
For each supply point i, the sum of the number of fast
charging stations (num_fcs_i) and slow charging stations
(num_scs_i) must be less than or equal to the total number of
parking spots (num_parking_spots_i) available at that supply
point.
For each supply point i, the total demand supplied
(demand_supplied_i) must not exceed the maximum possible
supply (max_supply_i) at that supply point. G. Result:
E. The cost function: The outcome of our optimization model is the calculation
of the most efficient distribution of the demand and supply
The function that determines the cost involves the matrix along with the ideal number of slow and fast charging
addition of the cost of dissatisfaction, 25 times the cost of stations, depicted in figure 3. Figure 4 presents the
demand mismatch, and 600 times the cost of infrastructure. performance details and figure 5 depicts the objective cost
The aim is to minimize this cost function. Therefore, the and accuracy score of the proposed methodology. Work done
optimization model takes the shape of the equation below: is an effort to identify the areas with the greatest potential for
installing EV charging points and suggest the fastest and most
𝑚𝑖𝑛 𝑆𝐶𝑆𝑗 ,𝐹𝐶𝑆𝑗 ,𝐷𝑖𝑗 ((𝐷𝑆𝑖𝑗 ∗ 𝐷𝑖𝑠𝑡𝑖𝑗 ) affordable charging options. It evaluates current EV adoption
trends and utilizes curve-fitting to predict future demand
suited for EV charging facilities. It then leverages the Gurobi
optimization solver to address cost minimization by
+ 25. (∑ 𝑎𝑏𝑠(𝐷𝑝𝑟𝑒𝑑,𝑖 − 𝐷𝑖 )) allocating demand to charging stations strategically. The
𝑖 results demonstrate the effectiveness of this approach in
optimizing EV charging station deployment, achieving an
+ 600 (∑(𝑆𝐶𝑆𝑗 + 1.5 ∗ 𝐹𝐶𝑆𝑗 ))) accuracy of over 93%. Notably, Scenario 1 achieved an
𝑗
optimal solution with a cost of INR 3,978,066/-, highlighting
the potential cost savings this method offers.

15th ICCCNT IEEE Conference,

Juneon
Authorized licensed use limited to: New York University. Downloaded 24-28, 2024, 07,2024 at 01:29:39 UTC from IEEE Xplore. Restrictions apply.
November
IIT - Mandi, Kamand, India
 IEEE - 61001

grid capacity, and electricity prices. This can help

balance the load on the grid and reduce overall
energy costs. [16]
ii. Renewable Energy Integration: As renewable
energy sources like solar and wind power become
more prevalent, EV charging stations may
increasingly rely on clean energy sources for
electricity. Integrating renewable energy generation
with charging infrastructure can further reduce the
environmental footprint of electric vehicles and
contribute to decarbonizing transportation. [17]
iii. Fast Charging Innovations: Research and
development efforts are likely to focus on
Fig. 3. Demand Allocation.
improving the speed and efficiency of fast charging
Performance details show that total time taken was 51.1 technologies. Innovations such as ultra-fast
seconds thus further reducing the latency. Optimizing the charging and wireless charging could become more
deployment of EV charging stations is essential to ensure a widespread, enabling EV drivers to recharge their
smooth shift to ecofriendly transportation and satisfy the vehicles quickly and conveniently, further reducing
rising demand for it. Table 1 gives a comparative analysis of range anxiety and increasing the appeal of electric
countries with the minimum number of individuals per single vehicles [18,19].
charging station, proposed methodology provides a technique iv. Dynamic Deployment Strategies: With the growth
that can further reduce the number of charging stations of EV adoption and changing mobility patterns,
required per area.
there may be a shift towards more dynamic
deployment strategies for charging infrastructure.
This could involve real-time monitoring of EV
usage and demand patterns to identify optimal
locations for new charging stations and adjust
existing infrastructure as needed. [20]
v. Collaborative Partnerships: Collaboration between
governments, utilities, automakers, and technology
companies will be crucial for scaling up EV
charging infrastructure efficiently. Public-private
Fig. 4. Performance Details.
partnerships can help address funding challenges,
streamline regulatory processes, and accelerate the
deployment of charging stations, particularly in
underserved areas. [21]
vi. Data-Driven Decision Making: Continued
advancements in data analytics and modeling
techniques will enable more accurate forecasting of
EV charging demand and optimization of charging
station placement. By leveraging data from EV
usage patterns, traffic flows, and demographic
trends, stakeholders can make informed decisions
to ensure that charging infrastructure meets the
needs of drivers effectively [22].
vii. Inclusive Access and Equity: Ensuring equitable
access to EV charging infrastructure will be
Fig. 5. Result.
essential for promoting widespread adoption of
VI. FUTURE OF OPTIMIZING THE DEPLOYMENT OF ELECTRIC electric vehicles. Efforts to address charging deserts
VEHICLE (EV) CHARGING STATIONS in underserved communities and provide charging
options for apartment dwellers and those without
The future of optimizing the deployment of electric
access to home charging will become increasingly
vehicle (EV) charging stations appears promising, with
important. [23]
continued advancements in technology and increasing
societal emphasis on sustainability. Here are some potential VII. CONCLUSION
future developments in this area:
The research investigates an algorithm designed to
i. Smart Charging Infrastructure: Future EV charging simulate the optimal placement of charging stations within a
stations are likely to incorporate smart technology, specific environment, with a primary goal of minimizing
allowing for better integration with the power grid construction costs. A comprehensive total social cost model
and more efficient charging processes. Smart is developed, encompassing both economic and
charging stations can communicate with vehicles to environmental factors. Economic costs, including
optimize charging times based on energy demand, construction and operational expenses, are considered

15th ICCCNT IEEE Conference,

Juneon
Authorized licensed use limited to: New York University. Downloaded 24-28, 2024, 07,2024 at 01:29:39 UTC from IEEE Xplore. Restrictions apply.
November
IIT - Mandi, Kamand, India
 IEEE - 61001

alongside environmental costs, such as dissatisfaction distribution network and traffic condition,” Energy Reports, vol. 9, pp.
stemming from carbon dioxide emissions or other 5246–5268, Dec. 2023, doi: 10.1016/j.egyr.2023.04.355.
environmental concerns faced by consumers. The model's [8] I. T. Gulbahar, M. Sutcu, A. Almomany, and B. S. K. K. Ibrahim,
“Optimizing electric vehicle charging station location on highways: A
constraints encompass maximum charging supply, vehicle decision model for meeting intercity travel demand,” Sustainability,
charging options, and charging distance. vol. 15, no. 24, pp. 16716, Dec. 2023, doi: 10.3390/su152416716.
Utilizing a genetic algorithm model, the study focuses on [9] M. B. Şenol, “A mixed integer programming (MIP) model for
evaluating navigation and task planning of human–robot interactions
a sample environment consisting of '64x64 blocks' to simulate (HRI),” Intelligent Service Robotics, vol. 12, no. 3, pp. 231–242, Mar.
the optimal distribution of 100 charging stations. Moreover, 2019, doi: 10.1007/s11370-019-00275-w.
sensitivity analysis is conducted to assess factors influencing [10] Zhang, H., Huang, Y., Liu, S., & Wu, T. “A mixed-integer linear
the total cost of charging station construction, revealing programming approach for electric vehicle charging infrastructure
planning. Transportation Research”, Part C: Emerging Technologies,
sensitivity to the number of charging stations, intersection vol. 102,pp. 1-15,2019.
charging demand, and daily charging probability. The [11] L. P. Karmali, A. Gholami, and N. Nezamoddini, “Integrated
contributions of the paper are multi-faceted: firstly, it optimization of production planning and electric trucks charging and
constructs a theoretical model for optimal charging station discharging scheduling,” Sustainable Energy, Grids and Networks, p.
distribution based on total social cost, providing a 101397, Apr. 2024, doi: 10.1016/j.segan.2024.101397.
foundational framework for simulating optimal distributions. [12] Patel, M. K., & Rahman, S..” Electric vehicle (EV) charging and
infrastructure: a review on models, technologies, and impacts”,
Secondly, a recognized genetic algorithm is employed to Renewable and Sustainable Energy Reviews, vol.82, p.4765-4789,
iteratively optimize charging station layouts. Thirdly, the use 2019.
of traditional Euclidean distance in parameter setting [13] H. Altaleb and Z. Rajnai, “Electric vehicle charging infrastructure and
enhances model alignment with actual demand patterns. charging technologies,” Haditechnika, vol. 54, no. 4, pp. 8–12, Jan.
2020, doi: 10.23713/ht.54.4.03.
Lastly, the introduction of a coefficient to account for
[14] M. B. Şenol, “A mixed integer programming (MIP) model for
charging station demand better reflects the actual distribution evaluating navigation and task planning of human–robot interactions
of EV charging demand within cities, facilitating (HRI),” Intelligent Service Robotics, vol. 12, no. 3, pp. 231–242, Mar.
determination of optimal station locations. results 2019, doi: 10.1007/s11370-019-00275-w.
demonstrate the effectiveness of this approach in optimizing [15] Zema, Tomasz, Adam Sulich, and Sebastian Grzesiak, "Charging
EV charging station deployment, achieving an accuracy of Stations and Electromobility Development: A Cross-Country
Comparative Analysis", Energies 16, no. 1: 32., 2023
over 93%. Notably, Scenario 1 achieved an optimal solution https://doi.org/10.3390/en16010032
with a cost of INR 3,978,066/- with an accuracy score of 96%, [16] O. Q. Wu, Ş. Yücel, and Y. H. Zhou, “Smart charging of electric
highlighting the potential cost savings this method offers. vehicles,” Social Science Research Network, Jan. 2019, doi:
Optimizing the deployment of EV charging stations is 10.2139/ssrn.3479455.
essential to ensure a smooth shift to ecofriendly transportation [17] P. Barman et al., “Renewable energy integration with electric vehicle
and satisfy the rising demand for it. technology: A review of the existing smart charging approaches,”
Renewable & Sustainable Energy Reviews, vol. 183, p. 113518, Sep.
2023, doi: 10.1016/j.rser.2023.113518.
REFERENCES [18] M. Brenna, F. Foiadelli, C. Leone, and M. Longo, “Electric vehicles
[1] O. Hortay and T. Pálvölgyi, “Driving forces in carbon dioxide charging technology review and optimal size estimation,” Journal of
emissions of the Hungarian transport sector,” Periodica Polytechnica. Electrical Engineering & Technology/Journal of Electrical
Transportation Engineering, vol. 50, no. 1, pp. 23–27, Dec. 2021, doi: Engineering & Technology, vol. 15, no. 6, pp. 2539–2552, Oct. 2020,
10.3311/pptr.15823. doi: 10.1007/s42835-020-00547-x.
[2] X. Ouyang and B. Lin, “An analysis of the driving forces of energy- [19] M. A. Ravindran et al., “A Novel technological Review on fast
related carbon dioxide emissions in China’s industrial sector,” charging infrastructure for electrical Vehicles: challenges, solutions,
Renewable & Sustainable Energy Reviews, vol. 45, pp. 838–849, May and future research directions,” Alexandria Engineering Journal
2015, doi: 10.1016/j.rser.2015.02.030. /Alexandria Engineering Journal, vol. 82, pp. 260–290, Nov. 2023,
[3] M. S. Mastoi et al., “An in-depth analysis of electric vehicle charging doi: 10.1016/j.aej.2023.10.009.
station infrastructure, policy implications, and future trends,” Energy [20] H. Wang, Q. Meng, and L. Xiao, “Electric-vehicle charging facility
Reports, vol. 8, pp. 11504–11529, Nov. 2022, doi: deployment models for dense-city residential carparks considering
10.1016/j.egyr.2022.09.011. demand uncertainty and grid dynamics,” Transportation Research.
[4] K. Afridi, “The future of electric vehicle charging infrastructure,” Part C, Emerging Technologies, p. 104579, Apr. 2024, doi:
Nature Electronics., vol. 5, no. 2, pp. 62–64, Feb. 2022, doi: 10.1016/j.trc.2024.104579.
10.1038/s41928-022-00726-w. [21] T. Dokka, S. SenGupta, and A. Bhardwaj, “Public EV charging
[5] Zhou, K., & Wu, J. “. An integrated modeling framework for optimal infrastructure - why charging behaviours matter for placement,
design of public electric vehicle charging infrastructure with ownership and operations?,” Social Science Research Network, Jan.
renewable energy integration”, Applied Energy, 236, 679-692,2019. 2022, doi: 10.2139/ssrn.4256502.
[6] M. M. Vazifeh, H. Zhang, P. Santi, and C. Ratti, “Optimizing the [22] F. Farhadi, S. Wang, R. Palacin, and P. Blythe, “Data-driven multi-
deployment of electric vehicle charging stations using pervasive objective optimization for electric vehicle charging infrastructure,”
mobility data,” Transportation Research. Part a, Policy and Practice, iScience, vol. 26, no. 10, p. 107737, Oct. 2023, doi:
vol. 121, pp. 75–91, Mar. 2019, doi: 10.1016/j.tra.2019.01.002. 10.1016/j.isci.2023.107737.
[7] M. Campaña and E. Inga, “Optimal deployment of fast-charging [23] C.-W. Hsu and K. Fingerman, “Public electric vehicle charger access
stations for electric vehicles considering the sizing of the electrical disparities across race and income in California,” Transport Policy,
vol. 100, pp. 59–67, Jan. 2021, doi: 10.1016/j.tranpol.2020.10.003.

15th ICCCNT IEEE Conference,

Juneon
Authorized licensed use limited to: New York University. Downloaded 24-28, 2024, 07,2024 at 01:29:39 UTC from IEEE Xplore. Restrictions apply.
November
IIT - Mandi, Kamand, India