` Selva: Publications
Selva Nadarajah [selvan@uic.edu]

My research addresses challenges at the interface of operations and finance arising in the energy industry using reinforcement learning and optimization. I am in particular interested in the operations of commodity and energy conversion assets (e.g., storage, transport, and production), their valuation and risk management, the role of corporations in the clean energy transition, and the social impact of this transition.

Influenced by my energy research and work with industry, I am passionate about making reinforcement learning and optimization methods easier to adopt by reducing the effort needed to deploy them effectively across different applications. To this end, I develop methods that embed solve intelligence and are capable of adapting over time and to instance-specific information in a problem class.

(click on paper title to view abstract)

Submitted Journal Papers

Energy and the Environment
  1. Decarbonizing buildings via energy demand response and deep reinforcement learning: The deployment value of supervisory planning and guardrails. (with D. Jang, L. Spangher, C. Spanos). [pdf]

  2. A review of the operations literature on real options in energy. (with N. Secomandi). [pdf; xlsx table; Invited review article; Under second round review at European Journal of Operational Research]

  3. Least squares Monte Carlo and pathwise optimization for merchant energy production. (with B. Yang, N. Secomandi). [ssrn; Under revision for third round review at Operations Research]
Reinforcement Learning and Optimization
  1. Self-adapting network relaxations for weakly coupled Markov decision processes. (with A. Cire). [pdf; Under revision for second round review at Management Science]

  2. Self-adapting robustness in demand learning. (with B. Chen, P. Pakiman, S. Jasin). [pdf; Under revision for resubmission to Operations Research]

  3. Self-guided approximate linear programs. (with P. Pakiman, N. Soheili, Q. Lin). [pdf; Under revision for third round review at Management Science]

Accepted/Published Journal Papers

Energy and the Environment
  1. Data-driven storage operations: Cross-commodity backtest and structured policies. (with C. Mandl, S. Minner, N. Gavirneni). [pdf; Forthcoming at Production and Operations Management]

  2. Meeting corporate renewable power targets. (with A. Trivella, D. Mohseni-Taheri). [pdf; video; Received the 2021 Commodity and Energy Markets Association Best Paper Award and the 2020 INFORMS ENRE Young Researcher Prize; Forthcoming at Management Science]

  3. Socially responsible merchant operations: Comparison of shutdown-averse CVaR and anticipated regret policies. (with A. Trivella). Operations Research Letters, 49(4), 2021. [pdf]

  4. Managing shutdown decisions in merchant commodity and energy production: A social commerce perspective (with A. Trivella, S.E. Fleten, D. Mazieres, D. Pisinger). Manufacturing and Service Operations Management, 23 (2), 2021. [pdf]

  5. Merchant energy trading in a network (with N. Secomandi). Operations Research, 66(5), 2018. [pdf]

  6. Comparison of least squares Monte Carlo methods with applications to energy real options (with F. Margot and N. Secomandi). European Journal of Operational Research, 256(1), 2017. [pdf]

  7. Relaxations of approximate linear programs for the real option management of commodity storage (with F. Margot and N. Secomandi). Management Science, 61(12), 2015. [pdf]
Reinforcement Learning and Optimization
  1. A data efficient and feasible level set method for stochastic convex optimization with expectation constraints. (with Q. Lin, N. Soheili, T. Yang). Journal of Machine Learning Research, 21(143), 2020. [pdf]

  2. Network-based approximate linear programming for discrete optimization (with A. Cire). Operations Research, 68(6), 2020. [pdf]

  3. Partial hyperplane activation for generalized intersection cuts (with A. Kazachkov, E. Balas, and F. Margot). Mathematical Programming Computation, 12, 2020. [pdf; Received the Tepper Egon Balas PhD student paper award]

  4. Revisiting approximate linear programming: Constraint-violation learning with applications to inventory control and energy storage (with Q. Lin and N. Soheili). Management Science, 66(4), 2020. [pdf]

  5. A level-set method for convex optimization with a feasible solution path (with Q. Lin and N. Soheili). SIAM Journal on Optimization, 28(4), 2018. [pdf]

  6. Relationship between least squares Monte Carlo and approximate linear programming (with N. Secomandi). Operations Research Letters, 45(5), 2017. [pdf]

  7. Less-Than-Truckload carrier collaboration problem: modeling framework and solution approach (with J. H. Bookbinder). Journal of Heuristics, 19(6), 2013. [pdf]

Conference Proceedings and Workshop Papers
  1. Offline-online reinforcement learning for energy pricing in office demand response: Lowering energy and data costs (with D. Jang, L. Spangher, T. Srivistava, M. Khattar, U. Agwan, and C. Spanos). Proceedings of the 8th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation (BuildSys '21), 2021. [pdf]

  2. A machine learning approach to methane emmisions mitigation in the oil and gas industry (with Jiayang Wang, Jingfan Wang, and P. Ravikumar). Tackling climate change with machine learning workshop, NeurIPS 2020. [pdf; Selected for a Spot Light Talk and Overall Best Paper. Research covered by Fortune magazine in its Eye on A.I. newsletter.]

  3. Self-guided approximate linear programs (with P. Pakiman, N. Soheili, and Q. Lin). Self-supervised learning workshop, NeurIPS 2020. [pdf]

  4. Least squares Monte Carlo and approximate linear programming: Error bounds and energy real option application (with N. Secomandi). Advances in supply chain finance and FinTech innovations, Foundations and Trends in Technology, Information and Operations Management, Now Publishers, 2020. [pdf]

  5. Interpretable user models via decision-rule Gaussian processes. (with D. Mohseni-Taheri, T. Tulabandhula). Advances in Approximate Bayesian Inference workshop, NeurIPS 2019. [pdf]

  6. Robust demand learning. (with B. Chen, S. Jasin). Workshop on Safety and Robustness in Decision Making, NuerIPS 2019.

  7. SMOILE: Shopper marketing optimization and inverse learning engine. (with A. Chenreddy, P. Pakiman, R. Chandrasekaran, R. Abens). Proceedings of the 25th ACM SIGKDD conference on knowledge discovery and data mining, Anchorage, Alaska, 2019. (accepted for oral presentation; acceptance rate 6.4%) [pdf]

  8. A three-echelon integrated production-distribution system (with J. H. Bookbinder). International Conference on Decision Sciences and Technology for Globalization, Decision Sciences Institute, Ghaziabad, India. 2008.

  9. Enhancing transportation efficiencies through carrier collaboration (with J. H. Bookbinder). BPC World Conference , Mumbai, India. 2007.

  10. Non-destructive evaluation by low pulsing acoustic tap technique: Spectral relationships (with T.S. Niranjan and A.V.Varun). Flight 2006, National Aerospace Symposium, Chennai, India. 2006.

Book Chapters
  1. Real option management of hydrocarbon cracking operations (with N. Secomandi, G. Sowers, and J. Wassick). Real Options in Energy and Commodity Markets, Now Publishers, 2017. [pdf]

Technical Reports
  1. Least squares Monte Carlo and approximate linear programming: Error bounds and energy real option application (with N. Secomandi). [pdf; extended version of the conference proceedings with the same title.

  2. Dynamic pricing for hotel rooms when customers request multiple-day stays (with Y. F. Lim, Q. Ding). [pdf]