Deploying an Intelligent Online Food Ordering System to Optimize

College Canteen Operations

Pranav Bari, Yuvraj Patil, Neerajkumar Sathawne

Tushar Suroshe and Atharva Patil

Department of Computer Science & Business Systems, JSPM’s Rajarshi Shahu College of Engineering, Pune,

Maharashtra, India

Keywords: College Canteen Management, Online Food Ordering System, Digital Transformation, MERN Stack, Real-

Time Order Processing, Machine Learning in Food Services, User Experience, Automated Payment

Solutions, Campus Dining Efficiency.

Abstract: College canteens often face issues like long lines, slow service, and crowded spaces, especially during peak

hours. To address these challenges, this paper explores a practical approach to building an online food

ordering system specifically designed for college campuses. Our solution uses familiar web and mobile tools,

combined with modern technologies like the MERN stack (MongoDB, Express, React, Node.js) and simple

machine learning techniques, to make ordering food easier and faster for students and staff. Key features

include a web application where users can place orders, pay digitally, and get personalized recommendations

based on past choices. This system not only reduces wait times but also helps canteen staff manage orders

more efficiently. Testing shows that the new system significantly improves order accuracy, speeds up the

process, and creates a more enjoyable experience for users. This paper provides a practical blueprint for

colleges looking to upgrade their canteen services and keep pace with students’ expectations in today’s digital

world.

1 INTRODUCTION

For students and staff on college campuses, lunchtime

can be a hectic experience. With everyone trying to

grab a meal during short breaks, college canteens

often become crowded, leading to long lines, delayed

service, and occasional order mix-ups. These issues

not only disrupt the flow of a busy day but also affect

the overall satisfaction and productivity of students

and staff.

In recent years, digital solutions have transformed

many industries, including the food service sector.

Online food ordering apps have become a part of

everyday life, offering convenience and saving time.

Inspired by these advancements, we see an

opportunity to bring similar benefits to college

canteens by implementing an online ordering system

tailored for campus use. This system would allow

students and staff to browse the menu, place their

orders ahead of time, pay digitally, and simply pick

up their meals when ready, skipping the usual wait in

line.

This paper presents the design and

implementation of an online ordering system built

specifically for college canteens. We explore how the

MERN stack (MongoDB, Express, React, Node.js)

provides a flexible and scalable foundation for

creating a system that can handle high traffic,

especially during peak hours. Additionally, by

incorporating simple machine learning models, the

system can suggest personalized menu options based

on a user’s order history, making it easier for them to

choose meals they enjoy.

Beyond convenience for users, this system also

benefits canteen staff by streamlining order

management. Orders come through in real-time,

minimizing miscommunication, and allowing staff to

focus on preparing food rather than handling long

queues. In this way, the system is not just a digital

replacement for manual processes—it’s a tool to

improve overall efficiency and create a smoother

dining experience. In the following sections, we’ll

delve into the specifics of the system’s design, the

technologies used, and how this approach can be

Bari, P., Patil, Y., Sathawane, N., Suroshe, T. and Patil, A.

Deploying an Intelligent Online Food Ordering System to Optimize College Canteen Operations.

DOI: 10.5220/0013613700004664

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 3rd International Conference on Futuristic Technology (INCOFT 2025) - Volume 3, pages 263-268

ISBN: 978-989-758-763-4

263

adapted for other campuses looking to modernize

their canteen services.

2 RELATED WORK

The Local Food Delivery System (Shah, Shah, et al. ,

2021) connects local vendors, such as street hawkers

and grocery stores, directly with customers,

addressing the demand for convenient food delivery

in urban areas. The application allows users to browse

local menus, place orders, and track their status in

real-time, enhancing user experience while

empowering local businesses. Additionally, the

system offers health-conscious food suggestions for

users with specific dietary needs, promoting healthier

eating habits . Built on ASP.NET and MS SQL

Server, it ensures robust data management and

integrates payment processing and location tracking

for efficient deliveries . Overall, this system not only

improves food delivery services but also supports the

economic empowerment of local vendors.

The Automated Canteen Ordering System by

Sharma et al. (2022) (Pandey, Sharma, et al. , 2022)

exemplifies advancements in food service

automation, addressing inefficiencies in traditional

ordering methods. This system allows users to place

orders online via an interactive e-menu, significantly

reducing wait times and enhancing customer

satisfaction. Utilizing modern web development

standards, the system features a user-friendly

interface and a robust backend for efficient order

processing and inventory management. By

automating the ordering process, it minimizes human

error and alleviates staff workload, leading to

improved operational efficiency. While designed for

canteens, the system's principles can be adapted for

various food service environments. Future

enhancements could include AI-driven

recommendations and integrated payment gateways.

This work serves as a foundational reference for our

implementation, highlighting the transformative

potential of technology in food service operations.

(Masurah, 2021) The development of

recommendation systems has gained significant

attention, particularly in the context of food ordering

applications for school students. Previous studies

have highlighted the importance of integrating

nutritional information with user preferences to

enhance the effectiveness of food recommendation

systems. For instance, the Mobile School Canteen

Food Ordering System proposes a solution that not

only facilitates food selection but also recommends

healthy meal options based on dietary criteria,

addressing the challenges faced by students during

school breaks.

This research (Mohamad, 2021) builds upon

existing literature in restaurant recommendation

systems, particularly those utilizing collaborative

filtering and content-based filtering techniques.

Notably, the proposed system distinguishes itself by

leveraging the number of orders as a primary input, a

novel approach not previously explored in the field.

The methodology incorporates k-nearest neighbor

techniques to identify similar buying patterns among

clients, enhancing the accuracy of recommendations

and addressing the challenges posed by sparse

datasets and evolving consumer preferences.

The proposed "Online Canteen Food Order

System" (Kogta, 2020) aims to enhance efficiency in

college cafeterias by automating the ordering process,

which traditionally relies on paper-based systems.

Previous studies have highlighted the limitations of

such systems, including time-consuming billing

processes and the risk of lost records. Innovations in

mobile technology and cloud-based applications have

been explored to improve customer experience and

streamline operations, emphasizing the need for a

more effective communication channel between

consumers and food service providers.

The "Canteen Food Ordering and Managing

System" (Antony, 2022) by Chanchal Antony et al.

(2022) presents a comprehensive digital solution for

canteen operations, addressing common issues such

as long queues and manual paperwork. The system

allows users to create accounts, browse a digital

menu, place orders, and make online payments. 1 It

also features a virtual queue to minimize wait times

and provides a platform for customer feedback. 2 The

admin interface includes functionalities for managing

food items, tracking orders, viewing transactions, and

generating reports. 3 This system aims to enhance

efficiency and customer satisfaction in canteen

services by leveraging digital technology.

A clustered database food recommendation

system (Shaikh, Shetgaonkar, et al. , 2019) utilizes K-

means clustering to speed up the process by grouping

similar items, which is particularly effective with

large datasets. An automated food ordering system

leverages an Android application to streamline order

placement and customization, using JAVA for the

front end and MySQL for the backend. The Zigbee-

based e-menu ordering system offers an affordable

solution for restaurants, featuring a user-friendly

graphical interface that simplifies order placement

and billing, even for illiterate users. Another proposed

system employs wireless touch-panel menus to

digitize the ordering process, reducing the need for

INCOFT 2025 - International Conference on Futuristic Technology

264

waitstaff and minimizing errors. This system uses

Zigbee for wireless communication and a PIC

microcontroller for menu coding. Additionally, an

automated food ordering system integrates data

mining algorithms like Apriori and K-means to

perform association mining and clustering, aiming to

reduce counter staff, eliminate calculation errors, and

manage queues efficiently. These systems

collectively highlight the potential of technology to

transform food service operations by improving

speed, accuracy, and user experience.

3 PROPOSED SYSTEM

The process begins at the Home page, which serves

as the main entry point for both users and admins.

From here, users can initiate the login process by

going to the User Login screen. Upon a successful

presumably lists various meal options or items

available for ordering.

After viewing the mess list, users can proceed to

the Menu for the particular mess or canteen. At this

point, the system checks the Authenticity of the

request, which could mean verifying the user's

identity or checking if the requested item is available.

If this check fails (No), the user may be redirected or

notified accordingly. But if everything is in order

(Yes), the system moves on to a more detailed

validation step involving Credit Points or Item

Availability. This step likely ensures that the user has

sufficient credits to place the order, or it verifies that

the requested item is still available in stock.

Once these verifications pass, users are given the

chance to add items to their Cart. After adding items,

a confirmation step appears, labeled as Confirmation,

where users review their orders before proceeding. If

users are satisfied and confirm the order (Yes), the

next step generates a Token for their transaction. This

token is crucial as it likely tracks the user's position in

a Queue, ensuring orders are processed in sequence.

Parallelly, there is an Order Page accessible from

the Queue, where the system performs another check

labeled Check User Eligibility. This might involve

double-checking user details, ensuring they meet

specific requirements, or perhaps confirming they

haven't exceeded any ordering limits. If a user fails

this eligibility check (No), the system notifies them

through Msg to User, potentially explaining the

reason for rejection. If eligible (Yes), the order gets

processed further.

From here, the order moves to the backend where

it is saved in the Database (DB). At this point, data is

gathered for ML Prediction and Data Analysis

processes. The ML Prediction step could use data

patterns to forecast future item availability or user

preferences, while Data Analysis could provide

insights for optimizing operations or managing stock.

Admins also have a role in this system, with a

separate Admin Login route from the Home page.

Upon logging in, admins gain access to the backend

tools, including an option to Update Availability of

Items. This feature allows them to adjust stock levels,

ensuring the menu reflects real-time availability.

Through this update mechanism, admins can

communicate changes in availability to users, which

can trigger the Through Msg to User step, updating

users on item status in case of shortages or other

adjustments.

In summary, outlines of system designed to

handle user authentication, item selection, order

validation, stock management, and data processing

for predictions and analysis. The structured

interaction between users and admins ensures smooth

operation, while backend processes enable better

management and insights for future improvements.

For this project, we’re using the MERN stack,

which includes MongoDB, Express, React, and

Node.js. This stack allows for a smooth flow of data

from the frontend to the backend, making it highly

efficient and suitable for building dynamic,

responsive web applications. MongoDB serves as our

database, storing user information, orders, item

availability, and other crucial data. Express is used for

routing and setting up the server-side logic, while

React is handling the user interface, allowing for

interactive and real-time updates. Node.js brings

everything together, providing the runtime for our

backend operations.

We also incorporated Nodemailer to handle OTP

(One-Time Password) verification via email, which is

essential for confirming user authenticity. When

users sign up or log in, they receive an OTP, which

they must enter to proceed, adding an extra layer of

security to the process.

For authentication, we're using JWT (JSON Web

Tokens). JWT tokens help verify users without

having to check their credentials every time they

make a request. When a user logs in successfully, a

JWT token is issued and sent to the frontend, where it

can be stored. For every subsequent request, this

token is passed back to the server for verification.

Lastly, role-based authorization is implemented

through Express middleware. Different roles like

User and Admin have different access privileges; for

example, only Admins can update item availability.

This role management setup ensures that only

Deploying an Intelligent Online Food Ordering System to Optimize College Canteen Operations

265

authorized users can perform certain actions, keeping

the system organized and secure.

Overall, the MERN stack, combined with tools

like Nodemailer and JWT-based authentication,

provides a reliable foundation for building a full-

featured application with enhanced security and

structured access control.

Figure 1: Order Management System Flowchart

3.1 Algorithm Used: Priority Queue

Algorithm

The Priority Queue algorithm is designed to help a

canteen process orders in the most efficient way

possible, balancing customer satisfaction and profit

while considering key operational constraints. It

prioritizes takeaway orders by calculating a "priority

score" based on several important factors: the profit

margin of each order, the time it takes to prepare, how

complex the order is, the customer’s preferences, and

how urgent the order is.

Here’s how each factor plays a role:

1. Profit Margin (P): This is how much

money the canteen makes from an order.

Orders that generate higher profits should be

given more attention.

2. Preparation Time (T): The time needed to

prepare the order. Shorter orders that can be

prepared quickly are given higher priority to

maximize efficiency during busy periods.

3. Order Complexity (C): This reflects how

difficult or time-consuming an order is to

prepare. Complex orders, which require

more ingredients or steps, are deprioritized

to keep the process smooth.

4. Customer Preference (CP): If a customer

has ordered a particular dish before or if it is

a popular dish, the system prioritizes that

order to enhance the customer’s experience

and satisfaction.

5. Urgency (U): This factor determines how

quickly the order needs to be completed.

Urgent orders, such as those that need to be

ready within a short time frame, are given

higher priority.

The algorithm calculates a priority score for each

order using the formula:

𝑷𝒓𝒊𝒐𝒓𝒊𝒕𝒚 𝑺𝒄𝒐𝒓𝒆 = (𝑷 / 𝑻) − 𝑪 + 𝑪𝑷 + 𝑼 (𝟏)

Where:

● P / T gives a higher priority to orders that

offer a good profit relative to the time

required to prepare them.

● C reduces the priority for complex orders.

● CP increases the priority for orders that

reflect customer preferences.

● U ensures that urgent orders are handled first.

Once each order has a priority score, the orders

are sorted from the highest to the lowest score,

ensuring that the most important and profitable orders

are handled first. The system also checks whether the

necessary ingredients are available, and if certain

items are out of stock, those orders are given lower

priority.

This algorithm allows the canteen to dynamically

manage its orders, ensuring that it maximizes profit,

keeps customers happy, and works within the

constraints of time, resources, and complexity.

3.2 Recommendation Algorithm Using

Cosine Similarity

The recommendation algorithm utilizes cosine

similarity to suggest items to a target user based on

their past interactions and the behaviors of similar

users. The process starts by calculating how similar

each user is to the target user using cosine similarity,

a metric that measures the angle between two vectors

of user-item interactions. This similarity score helps

identify users who have similar tastes or preferences

to the target user.

Once the similarity scores are computed, the

algorithm selects the most similar users and examines

the items they have interacted with, especially those

that the target user has not yet rated or purchased. By

aggregating the ratings or interactions from the

similar users, the algorithm generates a prediction

INCOFT 2025 - International Conference on Futuristic Technology

266

score for each item that the target user has not

interacted with.

Finally, the items are ranked based on these

scores, and the top recommendations are presented to

the target user. This method ensures that the

recommendations are personalized, taking into

account not just the target user's past behavior but also

the preferences of users who are most similar to them.

Input:

● A user-item interaction matrix M, where

rows correspond to users, columns

correspond to items, and the matrix elements

represent interaction values (e.g., ratings or

purchase frequency).

● The target user U

, for whom

recommendations are to be generated.

● The number of desired recommendations N.

Steps:

1. Calculate User Similarities:

For each user U

in the dataset:

○ Compute the similarity between U

and U

using the cosine similarity

formula:

𝑆𝑖𝑚𝑖𝑙𝑎𝑟𝑖𝑡𝑦 (𝑈



,𝑈



)

∑





𝑈



[𝑗] ⋅ 𝑈



[𝑗]



∑





𝑈



[𝑗]



⋅



∑





𝑈



[𝑗]



(2)

Rank Users by Similarity:

Arrange all users U

based on their

similarity scores in descending

order. Select the top K users who

are most similar to U

2. Aggregate Scores for Unrated Items:

For each item I

, that the target user U

has

not interacted with:

○ Compute a weighted score using

the interaction data from the top K

similar users:

𝑆𝑐𝑜𝑟𝑒 (𝐼



)=





(𝑆𝑖𝑚𝑖𝑙𝑎𝑟𝑖𝑡𝑦(𝑈



,𝑈



)

⋅ 𝑀[𝑈



,𝐼



](3)

○ Here, 𝑀[𝑈



,𝐼



] represents the

interaction value of user 𝑈



with

item 𝐼



3. Recommend Top Items:

Rank all items I

based on their computed

scores in descending order. Select the top N

items as recommendations for U

Output:

A list of N items recommended for the target user

based on their similarity with other users and

aggregated scores for unrated items.

4 EXPECTED OUTCOMES

By implementing this system, several key

improvements in user experience and operational

efficiency can be achieved:

Reduced Wait Times: The queue management

system, paired with real-time item availability

updates and user eligibility checks, helps streamline

the ordering process. With each order processed in

sequence, users experience shorter wait times,

especially during peak hours. The system also

minimizes bottlenecks by only allowing eligible users

to place orders, preventing unnecessary load and

delays.

Improved Order Accuracy: The system’s

layered verification process, including credit checks

and item availability confirmation, significantly

improves order accuracy. Each order goes through

multiple validation steps, ensuring that only items in

stock and available for the user’s credit level are

added to the cart. Additionally, the OTP verification

step prevents unauthorized access, reducing the

chances of errors due to fraudulent orders.

Enhanced User Experience: With the user-

friendly interface powered by the React frontend, and

a smooth backend flow managed by Node.js and

Express, users can navigate through the menu, add

items to the cart, and complete orders seamlessly. The

role-based access control ensures that users and

admins have customized, relevant options, which

further contributes to a streamlined experience.

Data-Driven Insights: By incorporating data

analysis and machine learning prediction models, the

system allows for better forecasting of user

preferences and inventory needs. This leads to

proactive stock management, where items can be

replenished before they run out, further enhancing

order accuracy and availability.

Deploying an Intelligent Online Food Ordering System to Optimize College Canteen Operations

267

5 CHALLENGES FACED

One of the primary challenges encountered in this

project involves integrating the payment gateway.

While the gateway facilitates smooth and secure

transactions, it also imposes a 2% brokerage fee on

each transaction. This brokerage can become costly

over time, especially with high transaction volumes,

reducing the overall profitability and increasing

operational expenses.

To address this, alternative solutions are being

considered to minimize these costs. Options might

include exploring payment providers with lower

transaction fees, integrating a direct bank transfer

system, or possibly incorporating a wallet or prepaid

credit system within the application. These

alternatives could help reduce dependency on

external gateways and improve cost efficiency.

However, each alternative requires careful

evaluation for security, user convenience, and

compatibility with existing infrastructure.

Balancing a seamless payment experience for

users while managing transaction fees remains a

complex challenge, and ongoing adjustments will be

essential to find the most cost-effective yet user-

friendly approach.

6 CONCLUSION

The implementation of the Intelligent Online Food

Ordering System for college canteens represents a

significant advancement in addressing the

challenges faced by students and staff during meal

times. By utilizing the MERN stack and integrating

machine learning techniques, we have developed a

robust platform that not only streamlines the

ordering process but also enhances user experience

through personalized recommendations and real-

time order management.

The system effectively reduces wait times,

minimizes order inaccuracies, and alleviates the

congestion typically associated with traditional

canteen operations. Furthermore, it empowers

canteen staff to focus on food preparation rather than

managing long queues, thereby improving overall

operational efficiency.

As we move forward, it is essential to continue

refining the system based on user feedback and

technological advancements to ensure it meets the

evolving needs of college communities. This

implementation serves as a practical blueprint for

other institutions looking to modernize their dining

services and align with the digital expectations of

today’s students. By embracing such innovations,

colleges can significantly enhance the dining

experience, ultimately contributing to the

satisfaction and productivity of their campus

populations.

REFERENCES

Shah, N., Shah, H., Sheth, S., Chauhan, R., & Desai, C.

(2021). Local food delivery system. SSRN Electronic

Journal. https://doi.org/10.2139/ssrn.3867478

Pandey, N., Sharma, S., Sharma, V., & Garg, T. (2022).

Automated Canteen Ordering System. International

Journal for Research in Applied Science and

Engineering Technology, 10(5), 865–870.

https://doi.org/10.22214/ijraset.2022.42356

Mohamad, Masurah. (2021). Mobile School Canteen Food

Ordering System. Mathematical Sciences and

Informatics Journal. 2. 102-110.

10.24191/mij.v2i2.16142.

Mohamad, M. (2021). Mobile school canteen food ordering

system. Mathematical Sciences and Informatics

Journal, 2(2), 102–110.

https://doi.org/10.24191/mij.v2i2.16142

Kogta, A. K. (2020). Cross platform application for canteen

food ordering system. International Journal of

Innovative Technology and Exploring Engineering,

9(8), 1005–1010.

https://doi.org/10.35940/ijitee.h6447.069820

Antony, C. (2022). Canteen food Ordering and Managing

system. International Journal of Current Science

Research and Review, 05(06).

https://doi.org/10.47191/ijcsrr/v5-i6-51

Shaikh, A. M., Shetgaonkar, A. C., Dalvi, S. S., Sawant, S.

S., Singh, K. V., Naik, S. K., & Gauns, R. K. (2019).

Food Ordering Management using Recommendations.

International Journal of Advanced Engineering

Research and Science, 6(6), 163–166.

https://doi.org/10.22161/ijaers.6.6.17

INCOFT 2025 - International Conference on Futuristic Technology

268