One of the major issues faced by the food industry, that has been accelerated particularly in the wake of a pandemic is food insecurity - ranging from unavailability of food material for the mass from panic buying to food wastage due to rotting in store shelves as well as owing to the former. The problem statement further defines the issues faced by a food delivery system, wherein they are forced to deal with a lot of perishable raw materials. The most important factor for such a company is to accurately forecast daily and weekly demand. Too much inventory in the warehouse means more risk of wastage, and not enough could lead to out-of-stocks - and push customers to seek solutions from your competitors. In such a scenario, the replenishment of majority of raw materials is done on weekly basis and since the raw material is perishable, the procurement planning is of utmost importance.
What it does
As a food warehouse optimizing system, our solution- SmarTangleWMS, would principally work to reduce and hence eliminate this issue. We have come up with the idea of a SaaS Warehouse Optimization System for a warehouse that covers tracking of the movement, quality and shelf life of the product and also predicts the requirement of specific products based on daily demands for long term usage so as to prioritize and deliver as per requirements and reduce food insecurity. We plan to implement our solution with the help of Tangle , ML and IoT for supply chain demand prediction and visualization, a foolproof food quality and movement tracking as well as to ensure legitimacy and convenience for customers. It works on predicting the supply demand gap as well as the seasonality of material sales to provide insights on factors such as shop rush hours and material necessity. The Tangle network by IOTA is a novel technology which is a strong competitor to the blockchain technology, if not a possible successor, especially in the field of supply chain. Furthermore, to increase competency in producing the best quality raw materials, we have used ML to implement a unique Quality Score to the raw materials provided by the suppliers based on the nutritional information of the material and its growth conditions. This warrants a clear cut competition among the suppliers to provide the finest raw material produce. Our solution has its paramount interest in integrating all the stakeholders involved in the system so as to predict the requirement of food and its trends, over a period of time. This would aid in a situation like the present one, where logistics has hit a wall due to the COVID pandemic while food insecurity and panic buying is on an all-time peak. Our proposed system would consist of a Django based Web application for the Warehouse & Store admins as well as a Flutter based mobile app for the customers ensuring an end-to-end solution.
How We built it
Our technical Stack is as Follows :
Machine Learning Model: ○ TensorFlow Keras API ○ FbProphet Model by Facebook (Time Series Model) ○ TPOT Regressor (Quality Score Prediction) ● Tangle Network ○ IOTA Devnet (public) Tangle Network ● Warehouse-End Web Application : ○ Django Framework (Back-End) ○ HTML, CSS, JS, Bootstrap (Front-End) ○ Docker & Kubernetes (Containerization) ○ Auth0 (Credential Validity) ○ Google Cloud Platform (Deployment)
● Customer-End Mobile Application : ○ Flutter (Android Development SDK using Dart) ○ Google Maps SDK (Maps and Location details) ● Prototyping: ○ Adobe XD ● Databases: ○ Influx DB from IBM Cloud Catalog (Processing of ML data) ○ Firebase (NoSQL : User data & Requests from the mobile application) ○ SQLite (SQL : Store Inventory & Raw Material information)
Challenges We ran into
We were relatively new to Tangle Networks, working with SaaS applications and implementation of ML forecasting alongside. We underwent extensive research, trial and error scenarios, etc to learn about various Time-Series forecasting models, and to select the one that suits our supply chain demands. We also ran into compatibility issues initially while working with the cloud and use of Auth0 in our application. Flutter was also new to a lot of us and we had frequently sun into roadblocks regarding integrations
Accomplishments that We're proud of
We're proud that we could finally implement a working prototype that aligns with our initial problem solution track.
What We learned
A lottt tbh, every challenge we faced added on. In general, it was our first project concerning a logistical domain, so we had to do a study on the workings and issues of a food warehouse and how an optimization system should assist. We learned about the workings of the novel Tangle Network and how it competes with Blockchain in the supply chain domain. We also progressed more towards time-series forecasting with FbProphet which was relatively new to us. Moreover, we experimented with Flutter to create an end-to-end application system as well
What's next for smarTangle Warehouse Optimisation System via Tangle Network
More integrations and stability !