AI for Sustainability

AI for Sustainability

By 2050, over 80% of the global population will live in cities while consuming food, water, and energy extracted far from the point of consumption. At present, California and Florida  produce approximately 75% of our nation’s fruit and vegetable supply, yet represent just 2% of the nation’s farmland. In California, this super concentration of fruits and vegetables and associated water withdrawals affects wildlife, poisoning the water supply with selenium and salts, and causing ground subsidence exceeding 28 feet in some areas. Adding to these concerns, approximately 5% of the natural gas supply (i.e., 2% of global energy production) is consumed to produce nitrogen fertilizers. Depletion of non-renewable phosphorus reserves has been addressed extensively. Around 2035, phosphorus is expected to peak, and its price has doubled in the past 10 years. Our food system may not be sustainable due to projected nutrient cost or scarcity, and the wide-spread eutrophication of our water supplies. However, emerging high-performance technologies (e.g., controlled environment agriculture, hydroponics, membrane filtration) could decouple production from geographic constraints by enabling production in locations or with alternative feedstocks not traditionally used in mainstream production.

The Institute team, with the support of two USDA projects, is designing, building and operating pilot-Testbeds to couple the water and nutrients in domestic wastewater with high-productivity Controlled Environment Agriculture (CEA) systems. Harnessing recent advances in wastewater treatment, food security, data analytics, and AI, the Institute is developing novel optimized technology-driven CEAs that can achieve high-areal vegetable productivity and increase food and nutritional security of urban communities with low operating cost and reduced energy consumption. 

Publications

Related work

  • Communication-Constrained Expansion Planning for Resilient Distribution Systems. Geunyeong Byeon, Pascal Van Hentenryck, Russell Bent, Harsha Nagarajan. INFORMS Journal on Computing, 32(4): 968-985, 2020.

  • Multi- disciplinary design optimization of distributed energy generation systems: The tradeoff between life cycle environmental and economic impacts. Junchen Yan, Osvaldo A. Broesicke, Xin Tong, Dong Wang, Duo Li, and John C. Crittenden. . Applied Energy, 2020 (to appear).

  • End-to-end game-focused learning of adversary behavior in security games.  Andrew Perrault, Bryan Wilder, Eric Ewing, Aditya Mate, Bistra Dilkina, and Milind Tambe. In AAAI Conference on Artificial Intelligence, pages 1378–1386, 2020.

  • Backwash sequence optimization of a pilot-scale ultrafiltration membrane system using data-driven modeling for parameter forecasting. B. Zhang, G. Kotsalis, J. Khan, Z. Xiong, T. Igou, G. Lan, and Y. Chen. Journal of Membrane Science, 612(15), 2020.

  • Parametric life cycle assessment for distributed combined cooling, heating and power integrated with solar energy and energy storage. Junchen Yan, Osvaldo A. Broesicke, Dong Wang, Duo Li, and John C. Crittenden.  Journal of Cleaner Production, 250:119483, 2020.

AI for Sustainability Team

Yongshen Chen
Yongsheng Chen
Georgia Institute of Technology
Leader
George Lan
George Lan
Georgia Institute of Technology
Co-leader
John Crittenden
John Crittenden
Georgia Institute of Technology
Bistra Dilkina
Bistra Dilkina
USC Site Director
University of Southern California
Pinar Keskinocak
Pinar Keskinocak
Georgia Institute of Technology
Pascal Van Hentenryck
Pascal Van Hentenryck
Director
Georgia Institute of Technology