ICT3909 Graduation Project: Building an AI-Powered Intelligent Meal Planning System to Combat Food Waste

Original Author/Maintainer: shappadappa Source Platform: GitHub Original Title: ICT3909-FYP Original Link: https://github.com/shappadappa/ICT3909-FYP Publication Date: 2026-05-28

This graduation project is an end-to-end AI application covering the entire process from dataset construction and model training to web application deployment. It aims to build an intelligent meal planning system using AI technology, generating personalized recipe recommendations based on users' dietary preferences, nutritional needs, and available ingredients to combat the global challenge of food waste.

Food waste is a severe global issue: according to the UN Food and Agriculture Organization (FAO), approximately one-third of food is wasted globally each year, while hundreds of millions of people face hunger. Household waste often stems from lack of planning (e.g., not knowing how to use ingredients, repeated purchases leading to expiration). This project attempts to solve this problem using AI by building an intelligent meal planning system to reduce waste.

Dataset Construction

Independently built a multi-dimensional dietary dataset containing recipe information, ingredient attributes, nutritional components, cooking time, etc., ensuring data quality and domain relevance.

Meal Planning Model

Developed multiple algorithms: rule-based recommendation system (filtering nutritional goals and dietary restrictions), collaborative filtering (learning preferences of similar users), content-based recommendation (analyzing similarity of ingredient and recipe features), and optimization algorithms (minimizing waste or cost while meeting nutritional constraints).

Web Application

Built an interactive web application where users input their ingredient list, dietary preferences, and nutritional goals, and the system generates personalized meal plans in real time.

Challenge 1: Data Scarcity

There are few public datasets in the meal planning domain; independent construction requires extensive collection, cleaning, and annotation. Solution: Adopt crowdsourcing or semi-automated data collection strategies, combined with manual verification to ensure quality, covering localized ingredients and recipes.

Challenge 2: Multi-Objective Optimization

Need to balance conflicting goals such as nutritional needs, taste preferences, waste reduction, cost control, and time constraints. Solution: Use multi-objective optimization algorithms or weighted scoring mechanisms, allowing users to adjust goal priorities and providing multiple alternative options.

Challenge 3: Balance Between Personalization and Generalization

Avoid information cocoons or lack of practical value. Solution: Introduce an exploration-exploitation trade-off mechanism, occasionally adding novel options while recommending preferred recipes.

1. End-to-end delivery: Covers the full AI project lifecycle from data collection, model training to web deployment, simulating real working environments.
2. Practical application scenarios: Directly addresses the social issue of food waste, reflecting the value of AI serving society.
3. Interactive demonstration: The web application allows users to try and evaluate, providing feedback channels for subsequent iterations.

1. Practice-oriented: Emphasize hands-on practice and end-to-end delivery, handling real data, solving real problems, and building usable products.
2. Interdisciplinary integration: Involves machine learning, software engineering, nutrition, user experience design, etc., fostering interdisciplinary thinking.
3. Social value: The topic focuses on Sustainable Development Goals (SDGs), emphasizing technical ethics and social responsibility.

• Mobile application: Expand to a mobile app for users to use while shopping or cooking.
• Image recognition: Integrate ingredient image recognition to automatically input ingredient lists via photos.
• Community features: Add a user community to support recipe sharing, recommendation evaluation, and experience exchange.
• Supermarket API integration: Connect to supermarket inventory and price data, considering ingredient availability and cost when making recommendations.

This project demonstrates solid technical capabilities and the awareness of using AI to solve practical problems. In today's era of severe food waste, the value of technology lies in bringing changes to humans and society.

For AI learners, this project serves as a reference example: how to choose meaningful topics, plan project scope, and transform algorithms into products. We look forward to more AI projects in the future that are both technically sound and socially valuable.