AI optimizes hog farming profitability: Advanced machine learning streamlines market-timing decisions

Deciding when to bring a hog to market has never been an easy task.

To maximize profits, farmers must gauge changing animal weights, pork prices, feed costs, and pen space, while also keeping an inventory of ready-for-market pigs to meet long-term contractual obligations with meat processors. With so many variables in flux, farmers face what scholars call the "curse of dimensionality"—essentially having too much data to best solve a problem analytically. Thus, optimal policies for such problems are essentially unobtainable.

Research co-authored by a UC Riverside School of Business professor, however, leverages artificial intelligence (AI) to break this curse. Using pricing and inventory data from a large hog farm operation in Illinois, the researchers developed a machine learning model to determine when, to whom, and how many pigs to sell to optimize long-term profits. The model recovered about 22% of the profits that farmers typically lose with traditional decision-making.

The paper is published in the SSRN Electronic Journal.

"Traditional farming methods tend to focus on immediate profits, overlooking how today's choices affect future earnings. This short-sightedness can mean missed opportunities and lower overall profits in the long run," said Danko Turcic, a UCR associate professor of operations and supply chain management and paper co-author.

Turcic explained that farmers face critical decisions once pigs reach the "finishing stage," an age of about six months when the hogs achieve a ready-for-market weight of about 200 pounds. It's a time when farmers must decide how many hogs to sell and how many to keep feeding to meet future contractual obligations or perhaps fetch higher profits from bigger pigs on the open market if they anticipate a rise in meat prices.

To help farmers with such decisions, Turcic and his co-authors started with an AI model commonly used in computer games and adapted it to the world of hog farming. This meant adding realistic constraints, like limits on how many hogs could be sold at once due to contracts with meatpackers, and ensuring the AI wouldn't try to sell more hogs than the farmer actually had.

Importantly, they made sure the AI's decision-making process was transparent and understandable. This "interpretability" is crucial in areas like farming and medicine, where users need to trust the AI's recommendations.

By understanding how the AI worked, the researchers discovered why it was so much better at hog selling than current practices. The AI could identify the best times to sell, like when market prices were high enough to offset the penalties for breaking contracts. It also strategically held back some hogs, anticipating either higher prices later or times when there might be fewer hogs available to sell.

While the study focuses on hog farming, its implications extend across industries involving perishable commodities and dynamic inventory management. Turcic suggests that similar systems could optimize decisions in agriculture, retail, and even product launch timing for consumer goods such as smartphones.

"Our work underscores the potential of AI not just to automate but to augment human decision-making with powerful tools that were previously unimaginable," Turcic said.

The study, "An Empirically Grounded Analytical Approach to Hog Farm Finishing Stage Management," is available online and has been accepted for publication in the Journal of Operations Management. Its co-authors are Panos Kouvelis of the Olin Business School at Washington University in St. Louis and Ye Liu of the Martin J. Whitman School of Management, Syracuse University, Syracuse, N.Y.

Provided by University of California - Riverside