The GRAMA Project – Enabling and Empowering Farmers through AI and Game Theory

The agriculture sector contributes to 16% of India’s GDP employing almost 50% of the total work-force. 85% of the farmers are small and marginal, holding less than 5 acres of land and they do not have access to modern technological advances in  agriculture.   Agriculture in India is largely dependent on nature; variable climate and more recently global warming make farming both challenging and volatile. Small and marginal farmers continue to be trapped in a vicious circle of low growth, low income, and high debt.

The GRAMA project stands for Game Theory, Random Processes, Artificial Intelligence, and Machine Learning for Agriculture.

We are working with NABARD (National Bank for Agriculture and Rural Development) and BEL (Bharat Electronics Limited) to develop and implement novel algorithms and principled solutions to help farmers. These algorithms and solutions  will be powered by tools from Artificial Intelligence and Game Theory.

The exhibit shows the problems that we are currently looking into.

Crop Planning  System (CROP-S)

A mismatch between the crops produced by farmers and the respective market demands could potentially lead to large-scale crop dumping.  This regularly leads to huge financial losses and distress to the farmers. To alleviate this problem, we address the macro-level problem of district level or taluka level agricultural

crop planning in any given state. Our interest is in how the Government or any state administration could make a principled recommendation on which crop acreages (number of acres cultivated under each crop) to allocate in which districts or geospatial regions in a given state, so as to match the demand for the crops and maximize the profits for the farmers. CROP-S  determines an assignment of crop

acreages to districts so as to maximize the profits for the farmers while simultaneously ensuring required crop security levels for each district. CROP-S uses data about predicted demands, transportation costs, compliance ratios (fraction of farmers who will follow the recommended crop plan), and historical data about yields and prices to arrive at an optimal allocation of crop acreages to districts.

Crop Price Prediction (PREPARE)

Accurate prediction of agricultural crop prices is a crucial input for decision-making by various stakeholders in agriculture: farmers, consumers, retailers, wholesalers, and the Government. These decisions have significant implications including, most importantly, the economic well-being of the farmers. In this project,  our objective is to accurately predict crop prices using historical price information, climate conditions, soil type, location, and other key determinants of crop prices. This is a technically challenging problem, which has been attempted before but much remains to be done.  We have  developed an innovative deep learning based approach to achieve increased accuracy in price prediction. The proposed approach uses graph neural networks (GNNs) in conjunction with a standard convolutional neural network (CNN) model to exploit geospatial dependencies in prices. Our approach works well with noisy legacy data and produces a performance that is at least 20% better than the results available in the literature. We are able to predict prices up to 30 days ahead. We have chosen two vegetables, potato (stable price behavior) and tomato (volatile price behavior) and worked with noisy public data available from Indian agricultural markets.

Crop Recommendation (ACRE)

A key challenge faced by small and marginal farmers is to determine which crops to grow to maximize their utililty. With a wrong choice of crops, farmers could end up with sub-optimal yields and low, and possibly even loss of, revenue. ACRE is a tool that implements a principled approach to choose a crop or a portfolio of crops, to maximize the utility to the farmer. ACRE uses available data such as soil  characteristics, weather conditions, and historical yield data, and uses state-of-the-art machine learning/deep learning models to compute an estimated utility to the farmer. ACRE’s recommendation system generates several portfolios of crops and ranking them. The technical novelty of ACRE is to use Sharpe Ratio, a risk metric used in financial portfolio selection to rank various alternate crop portfolios. We use publicly available data from agmarknet portal in India to perform several insightful experiments with ACRE.

Agri-Market Design (PROSPER)

PROSPER is short for Protocol for Selling Agricultural Produce for Enhanced Revenue.

The focus is on designing a mechanism for selling agricultural produce that is attractive to the farmers as well as the customers (buyers). We have started with a mechanism that uses an interesting Nash bargaining formulation which, for a two player seller-buyer game, provides a unique solution that satisfies properties that are desirable to the seller as well as the buyer. In our formulation, the seller is a farmer producer organisation (FPO) acting on behalf of all its farmer members while the second player is virtual aggregated player representing all buyers.

PROSPER first collects (a) the costs incurred (and honestly reported) by the farmers and (b) the strategic bids from the buyers. PROSPER then generates a revised bid vector using a Nash bargaining approach to ensure attractive utilities for farmers and buyers. The key component of PROSPER is a greedy version of the generalized second price (GSP) auction with reserve prices.

Cost Effective, Quality Assuring Procurement of Agricultural INPUTS (PROMISE)

Procuring agricultural inputs such as seeds, fertilizers, and pesticides, at desired quality levels and at affordable cost, forms a critical component of agricultural input operations. This is a particularly challenging problem being faced by small and marginal farmers in any emerging economy. Farmer collectives (FCs), which are cooperative societies of farmers, offer an excellent prospect for enabling cost-effective procurement of inputs with assured quality to the farmers. In this project, our objective is to design sound, explainable mechanisms by which an FC will be able to procure agri-inputs in bulk and distribute the inputs procured to the individual farmers who are members of the FC. In the methodology proposed, an FC engages qualified suppliers in a competitive, volume discount procurement auction in which the suppliers specify price discounts based on volumes supplied. The desiderata of properties for such an auction include: minimization of the total cost of procurement; incentive compatibility; individual rationality; fairness; and other business constraints. An auction satisfying all these properties is analytically infeasible and a key contribution of this paper is to develop a deep learning based approach to design such an auction. We have successfully experimented with  two realistic, stylized case studies from chili seeds procurement and a popular pesticide procurement to demonstrate the efficacy of these auctions.

AIML Pipeline for Agriculture Domain (AGRI-VAAHAN)

A wide range of issues in the digital agriculture space have been addressed using AIML techniques. Prominent problems in this space include crop price prediction, crop yield prediction, and crop recommendation. Keeping these problems in mind, we have built a platform, AGRI-VAAHAN which is an extensible AI-ML pipeline to seamlessly integrate the end-to-end activities in agriculture data analysis. Agri-Vaahan  is a publicly-accessible web application. It offers a wide variety of data imputation, data cleaning, and data analysis techniques which are especially useful in the context of agriculture data.

GRAMA Research Publications

  • Mayank Ratan Bhardwaj, Azal Fatima, Inavamsi Enaganti, and Y. Narahari. Incentive Compatible Mechanisms for Efficient Procurement of Agricultural Inputs for Farmers through Farmer Collectives. In ACM SIGCAS/SIGCHI Conference on Computing and Sustainable Societies (ACM COMPASS) 2022, pp. 696-700.
  • Rohit Patel, Inavamsi Enaganti, Mayank Ratan Bhardwaj, and Y. Narahari. A Data-driven, Farmer-oriented Agricultural Crop Recommendation Engine (ACRE). Proceedings of  International Conference on Big Data Analytics 2022,  Lecture Notes in Computer Science, Springer, pp. 227-248.
  • Mayank Ratan Bhardwaj, Bazil Ahmed, Prathik Diwakar, Ganesh Ghalme, and Y. Narahari. Designing Fair, Cost-optimal Auctions based on Deep Learning for Procuring Agricultural Inputs through Farmer Collectives. To appear:  Proceedings of 15th International Conference on Automation Science and Engineering (IEEE CASE 2023), Auckland, New Zealand.
  • Mayank Ratan Bhardwaj, Jaydeep Pawar, Abhijnya Bhat, Deepanshu, Inavamsi Enaganti, Kartik Sagar, and Y. Narahari. An innovative Deep Learning Based Approach for Accurate Agricultural Crop Price Prediction. To appear:  Proceedings of 15th International Conference on Automation Science and Engineering (IEEE CASE 2023), Auckland, New Zealand.
  • Mayank Ratan Bhardwaj, Abhishek Chaudhary, Inavamsi Enaganti, Kartik Sagar, and Y. Narahari. A Decision Support Tool for District Level Planning of Agricultural Crops for Maximizing Profits of Farmers. To appear: Proceedings of 15th International Conference on Automation Science and Engineering (IEEE CASE 2023), Auckland, New Zealand.

GRAMA Ph.D. Thesis

Mayank Ratan Bhardwaj. Novel Algorithms for Improving Agricultural Planning and Operations using Artificial Intelligence and Game Theory. July 2023 (Submitted).

GRAMA M.Tech. Projects (Completed)

  • Abhishek Kumar Chaudhary : CROP-S: A Decision Support Tool for Acreage Planning of Agricultural Crops for Maximizing Profits of Farmers. M.Tech. Project Thesis. June 2023
  • Chaitanya Chennam: PROSPER : Protocol for Optimal Selling of Agricultural Produce for Enhanced Revenue. M.Tech. Project Thesis. June 2023.
  • Kaushik Kukadia: Agri-Vaahan 2.0: AI-ML Pipeline for Agriculture. M.Tech. Project Thesis. June 2023.
  • Kishan Mittal: ACRE 2.0: Agricultural Crop Recommendation for Farmers. M.Tech. Project Thesis. June 2023.
  • Jaydeep Pawar: Using Geospatial Information and Improved Deep Learning Methods for Crop Price Prediction. M.Tech. Project Thesis. June 2023.
  • Sneha Negi: Designing Fair, Cost-optimal Combinatorial Auctions based on Deep Learning for Procuring Agricultural Inputs. M.Tech. Project Thesis. June 2023.
  • Meghnath Singh (M.Tech (AI) Student). Yield and Grade Prediction for Horticultural Crops using Visual Analytics.  M.Tech. Project Thesis. June 2022
  • Deepanshu (M.Tech (CSE) Student). Improved Methods for Agriculture Crop Price Prediction.M.Tech. Project Thesis. June 2022
  • Rohit Patel (M.Tech (CSE) Student). ACRE: Agriculture Crop Recommendation Engine.M.Tech. Project Thesis. June 2022
  • Azal Fatima (M.Tech. (CSE) Student). Mechanism Design For Efficient Procurement of Agricultural Inputs for Farmers.M.Tech. Project Thesis. June 2022
  • Sowjanya Pidathala (M.Tech. (AI) Student) and Vaidehi Bhaskara (Research Intern).  AGRI-VAHAN: An AIML Pipeline for Agriculture.M.Tech. Project Thesis. June 2022
  • Someshwar Arnoorla (M.Tech. (AI) Student). Agricultural Crop Price Prediction using Transformer Encoder Based Wide and dep Networks.M.Tech. Project Thesis. July  2021

GRAMA M.Tech. Projects (ongoing)

  • Y. Geetha Charan. Agri Vaahan 3.0.
  • Ankur Garg. Large Language Models in Digital Agriculture.
  • Kamble Sahil. Geneartive AI in Digital Agriculture.
  • Anand Nakat. Embedding Price Prediction into Crop Recommendation.
  • Prashanth Vithule. CROP-S 2.0.

GRAMA Interns

  • Sanath Patil. Ramaiah Institute of Technology. Price Prediction.
  • Abhijnya Bhat, PES University.  Price prediction and crop recommendation.
  • Vishisht Rao. PES University. Input Procurement Mechanisms and Marketplace for Selling Agricultural Produce.
  • Prathik Diawakar. UG Student, IISc. Input Procurement Mechanisms
  • Hritik Bana. IISER-Bhopal. Price prediction.

GRAMA Team