Jonathan Eckstein

E-mail: user name jeckstei host business domain suffix rutgers.edu

Voice: (848) 445-0510
Fax: (732) 445-1133

Webumé

I am a (tenured) professor in the MSIS (Management Science and Information Systems) department, Rutgers Business School, Rutgers University, with my office on the New Brunswick campus. I was department vice-chair July 2008 through June 2011. I am a member of RUTCOR.

In Fall 2014, I was elected a fellow of INFORMS.

Until going on sabbatical in the 2014-2015 academic year, I was coordinator of Rutgers' BAIT major. This major graduated its first students in 2014. RBGA named me "Best BAIT Professor" for the 2013-2014 academic year.

With William E.Hart and Cynthia Phillips, I won the 2006 INFORMS COIN-OR Cup.

I won the 2003 Thomas H. Mott, Jr. award for excellence in teaching, by vote of Rutgers' undergraduate business majors.

For the 2000-2001 and 2007-2008 academic years, I was on sabbatical at the ORFE Department at Princeton University.

Prior to Rutgers, I spent about four years as a research scientist at the now-defunct Thinking Machines Corporation. The firm produced parallel supercomputers and associated software. Much of its technology was acquired by Sun Microsystems and Oracle (which, very much later, merged together).

Prior to Thinking Machines, I taught for two years at Harvard Business School.

Before that, I obtained Masters (1986) and Ph.D. (1989) degrees in Operations Research from MIT, at the Operations Research Center. My dissertation advisor was Dimitri Bertsekas.

Before that, I worked as a programmer/analyst at Xenergy, Inc (since acquired by another firm). In particular, I was the lead developer for XenCAP, and responsible for the software design of RECAP. The firm described these software products as pivotal in Xenergy's history. Xenergy claims XenCAP has been used to perform energy audits of over two billion square feet of commercial building space, and RECAP has been used to energy-audit over 3.5 million homes.

My undegraduate degree was in Mathematics, from Harvard University (1980, Phi Beta Kappa, Magna Cum Laude)

I am a member of the The Mathemetical Programming Society, INFORMS and its Computing Society, and SIAM.

Grants, Projects, and Service, and Other Professional Activities

I have served several times on the board of the INFORMS Computing Society , and have been its representative to the INFORMS subdivision council.

Principal investigator of the following grants:

National Science Foundation grant CCR-9902092, in the Computer-Communications Research program of the "CISE" directorate, 1999-2004.

National Science Foundation grant CCF-1115638, in the Computer-Communications Foundations program of the "CISE" directorate, 2011-2015.

Air Force Office of Scientific Research grant FA9550-11-1-0164, Optimization and Discrete Mathematics Program, 2011-2015 (Andrzej Ruszczynski, co-PI).

Air Force Office of Scientific Research grant FA9550-15-1-0251, Optimization and Discrete Mathematics Program, 2015-2019 (Andrzej Ruszczynski, co-PI).

National Science Foundation grant CCF-1617617, in the Computer-Communications Research program of the "CISE" directorate, 2016-2020.

I wrote most of the PEBBL parallel branch-and-bound software package and am taking a leading role in its continued development. I also made contributions to ACRO (which included PEBBL), PICO (defunct), and UTILIB. This work is in collaboration with Sandia National Labs.

I designed the YASAI Monte Carlo simulation add-in for Microsoft Excel and have overseen its development, mostly by Rutgers undergraduate students.

On January 1, 2019, I became editor-in-chief of Mathematical Programming Computation. In the past I have served on the editorial boards of Computational Optimization and Applications and Optimization Methods and Software, and as a technical editor for Mathematical Programming Computation.

Open-Source Software Projects

PEBBL: C++ framework for building customized branch-and-bound algorithms, with flexible and scalable parallelism through MPI. Developed with Cythia A. Phillips and William E. Hart.
BranchAndBound.jl: An abstract branch-and-bound framework for Julia.
YASAI: Excel add-in for Monte Carlo simulation, written in VBA. Designed to be simple to use and suitable for teaching. Developed with numerous Rutgers undergraduate students.
projSplitFit: Python modules facilitating experimentation with projective splitting methods for convex data fitting problems. Developed with Patrick R. Johnstone.

Publications and Working Papers

Note: This list is fairly comprehensive but not completely exhaustive. I am the first or sole author unless otherwise indicated.

Stochastic Projective Splitting, with Patrick R. Johnstone (first author): Computational Optimization and Applications, appeared online September 2023.
Projective Hedging Algorithms for Multistage Stochastic Programming, Supporting Distributed and Asynchronous Implementation, with Jean-Paul Watson and David L. Woodruff,: Operations Research, appeared online July 2023. Earlier version available as Optimization Online working paper 2018-10-6895.
Projective Splitting with Forward Steps, with Patrick R. Johnstone (first author): Mathematical Programming 191(2):631-670 (2022).
A covering decomposition algorithm for power grid cyber-network segmentation, with Emma S. Johnson (first author), Santanu S. Dey, Cynthia A. Phillips, and John D. Siirola: Optimization Online preprint 2021-12-8725, December 2021.
Stochastic Projective Splitting: Solving Saddle-Point Problems with Multiple Regularizers, with Patrick R. Johnstone (first author), Thomas Flynn, and Shinjae Yoo: ArXiv preprint 2106.13067, June 2021.
Single-Forward-Step Projective Splitting: Exploiting Cocoercivity, with Patrick R. Johnstone (first author): Computational Optimization and Applications, 78(1):125-166 (2021).
Deriving Solution Value Bounds from the ADMM: Optimization Letters 14(6), 1289-1303 (2020).
Relative-error inertial-relaxed inexact versions of Douglas-Rachford and ADMM splitting algorithms, with M. Marques Alves (first author), Marina Geremia, and Jefferson G. Melo: Computational Optimization and Applications, 75(2):389-422 (2020).
Projective Splitting with Forward Steps Only Requires Continuity, with Patrick R. Johnstone (first author): Optimization Letters 14(1):229-247 (2020).
Convergence Rates for Projective Splitting, with Patrick R. Johnstone (first author): SIAM Journal on Optimization 29(3):1931-1957 (2019).
REPR: Rule-Enhanced Penalized Regression, with Ai Kagawa and Noam Goldberg: INFORMS Journal on Optimization 1(2):143-163 (2019).
Projective Splitting with Forward Steps: Asynchronous and Block-Iterative Operator Splitting, with Patrick R. Johnstone (first author): Optimization Online working paper 2018-03-6533 and ArXiv preprint 1803.07043, March 2018.
Relative-Error Approximate Versions of Douglas-Rachford Splitting and Special Cases of the ADMM, with Wang Yao: Mathematical Programming, 170(2), 417-444 (2018).
Asynchronous Block-Iterative Primal-Dual Decomposition Methods for Monotone Inclusions, with Patrick L. Combettes (first author): Mathematical Programming 168(1-2):645-672 (2018).
Introductory Relational Database Design for Business, with Microsoft Access, with Bonnie R. Schultz: Wiley, 2018 (student companion website and instructor companion website available from the publisher.
Rule-Enhanced Penalized Regression by Column Generation using Rectangular Maximum Agreement, with Noam Goldberg and Ai Kagawa: Proceedings of the 34th International Conference on Machine Learning (ICML 2017), appearing as Proceedings of Machine Learning Research 70:1059-1067 (2017).
Approximate ADMM Algorithms Derived from Lagrangian Splitting, with Wang Yao: Computational Optimization and Applications, 68(2):363-405 (2017).
A Simplified Form of Block-Iterative Operator Splitting, and an Asynchronous Algorithm Resembling the Multi-Block Alternating Direction Method of Multipliers.: Journal of Optimization Theory and Applications, 173(1):155-182 (2017). Corrections to one of the proofs are in a "forum" note in the same journal.
Approximate Versions of the Alternating Direction Method of Multipliers, with Wang Yao: Optimization Online working paper 2016-01-5276, January 2016.
Multilevel Optimization Modeling for Risk-Averse Stochastic Programming, with Deniz Eskandani and Jingnan Fan: INFORMS Journal on Computing, 28(1):112-128 (2016), plus online supplement.
PEBBL: An Object-Oriented Framework for Scalable Parallel Branch and Bound, with William E. Hart and Cythia A. Phillips: Mathematical Programming Computation, 7(4):429-429 (2015). Longer version available as RUTCOR Research Report RRR 9-2013, September 2013.
A Bundle Method for Exploiting Additive Structure in Difficult Optimization Problems, with Welington de Oliveira (lead author): E-print 2015-05-4935, Optimization Online, May 2015.
Object-Parallel Infrastructure for Implementing First-Order Methods, with an Example Application to LASSO, with György Mátyásfalvi: E-print 2015-01-4748, Optimization Online, January 2015.
Understanding the Convergence of the Alternating Direction Method of Multipliers: Theoretical and Computational Perspectives, with Wang Yao.: Pacific Journal on Optimization, 11(4):619-644 (2015). Expanded version available at Optimization Online.
PEBBL 1.4.1 User's Guide, with William E. Hart and Cythia A. Phillips: RUTCOR Research Report RRR 2-2014, September 2014.
Time Consistency Versus Law Invariance in Multistage Stochastic Optimization with Coherent Risk Measures: Multilevel Optimization Modeling and Computational Complexity: RUTCOR Research Report RRR 7-2013, August 2013.
Augmented Lagrangian and Alternating Direction Methods for Convex Optimization: A Tutorial and Some Illustrative Computational Results: RUTCOR Research Report RRR 32-2012, December 2012.
Complexity of Bilevel Coherent Risk Programming: RUTCOR Research Report RRR 17-2012, April 2012.
A Practical Relative Error Criterion for Augmented Lagrangians: with Paulo J.S. Silva. Mathematical Programming 141(1):319-348 (2013).
An Improved Branch-and-Bound Method for Maximum Monomial Agreement: with Noam Goldberg. INFORMS Journal on Computing, 24(2):328-341 (2012).
Sparse Weighted Voting Classifier Selection and its LP Relaxations: with Noam Goldberg (lead author). Information Processing Letters 112:481-486 (2012)
Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers: with Stephen Boyd (lead author), Neal Parikh, Eric Chu, and Borja Peleato. Foundations and Trends in Machine Learning, 3(1):1-122 (2011).
Boosting Classifiers with Tightened L₀-Relaxation Penalties: with Noam Goldberg. Proceedings of the the 27^th International Conference on Machine Learning (ICML 2010), 383-390 (2010).
Proximal Methods for Nonlinear Programming: Double Regularization and Inexact Subproblems: with Paulo J.S. Silva. Computational Optimization and Applications, 46(2):279-304 (2010).
General Projective Splitting Methods for Sums of Maximal Monotone Operators: with B.F. Svaiter, SIAM Journal on Control and Optimization, 48(2):787-811 (2009).
An Improved Branch-and-Bound Method for Maximum Monomial Agreement: with Noam Goldberg. OPT 2008 Optimization for Machine Learning Workshop, Neural Information Processing Systems Workshop (NIPS08), December 2008.
Optimal Information Monitoring under a Politeness Constraint: with Avigdor Gal and Sarit Reiner. INFORMS Journal on Computing 20(1):3-20 (2008).
Arrival Rate Approximation by Nonnegative Cubic Splines: with Farid Alizadeh (first author), Nilay Noyan, and Gabor Rudolf. Operations Research, 56(1):140-156 (2008).
A Family of Projective Splitting Methods for the Sum of Two Maximal Monotone Operators: with B.F. Svaiter. Mathematical Programming 111(1-2):173-199 (2008).
Pivot, Cut, and Dive: A Heuristic for Mixed 0-1 Integer Programming: with Mikhail Nediak. Journal of Heuristics 13(5):471-503 (2007).
Massively Parallel Mixed-Integer Programming: Algorithms and Applications: with William E.Hart and Cynthia Phillips. In Parallel Processing for Scientific Computing, M.A. Heroux, P. Raghavan, and H.D. Simon, editors, SIAM Books, November 2006.
PEBBL 1.0 User Guide: with Cynthia A. Phillips and William E. Hart. July 2006. RUTCOR Research Report RRR 19-2006, August 2006.
Double-Regularization Proximal Methods, with Complementarity Applications: with Paulo J.S. Silva (lead author). Computational Optimization and Applications 33(2-3):115-156 (2006).
Depth-Optimized Convexity Cuts: with Mikhail Nediak. Annals of Operations Research 139:95-129 (2005)
Scheduling of Data Transcription in Periodically Connected Databases: with Avigdor Gal (lead author) and Zachary Stoumbos, Stochastic Analysis and Applications 21(5):1021-1058 (2003).
A Practical General Approximation Criterion for Methods of Multipliers Based on Bregman Distances: Mathematical Programming 96(1):61-86 (2003).
The Maximum Box Problem and its Application to Data Analysis: with Peter L. Hammer, Ying Liu, Mikhail Nediak, and Bruno Simeone. Computational Optimization and Applications, 23(3):285-298 (2002).
YASAI: Yet Another Add-In for Teaching Elementary Monte Carlo Simulation in EXCEL: with Steven T. Riedmueller. INFORMS Transactions on Education 2(2), http://ite.pubs.informs.org/Vol2No2/EcksteinRiedmueller/.
Managing Periodically Updated Data in Relational Databases: A Stochastic Modeling Approach: with Avigdor Gal (lead author). Journal of the ACM, 46(6):1141-1183 (2001).
A letter to the editor published by the New York Times on December 10, 2000.
PICO: An Object-Oriented Framework for Parallel Branch and Bound: with Cynthia A. Phillips and William E. Hart. Proceedings of the Workshop on Inherently Parallel Algorithms in Optimization and Feasibility and their Applications, Studies in Computational Mathematics, Elsevier Scientific, 219-265 (2001). Older versions available as RUTCOR Research Report RRR 40-2000, August 2000, and report SAND2000-3000, Sandia National Laboratories, December 2000.
Rescaling and Stepsize Selection in Proximal Methods using Separable Generalized Distances: with Paulo J.S. Silva (lead author) and Carlos Humes Jr.. SIAM Journal on Optimization, 12(1):238-261 (2001).
Smooth Methods of Multipliers for Monotone Complementarity Problems: with M. C. Ferris. Mathematical Programming, 86(1):65-90 (1999).
Approximate Iterations in Bregman-Function-Based Proximal Algorithms: Mathematical Programming, 83(1) 113-123 (1998).
Operator Splitting Methods for Monotone Affine Variational Inequalities, with a Parallel Application to Optimal Control: with M. C. Ferris. INFORMS Journal on Computing, 10(2):218-235 (1998).
Resource Management in a Parallel Mixed Integer Programming Package: with William E. Hart and Cynthia A. Phillips. Proceedings of the Intel Supercomputer Users Group Conference, Albuquerque, NM, June 11-13, 1997.
How Much Communication Does Parallel Branch and Bound Need?: INFORMS Journal on Computing, 9(1):15-29 (1997).
Distributed versus Centralized Storage and Control for Parallel Branch and Bound: Mixed Integer Programming on the CM-5: Computational Optimization and Applications, 7(2):199-220 (1997).
Parallel Computing: Encyclopedia of Operations Research and Management Science, S. I. Gass and C. M. Harris, eds., Kluwer Academic, Boston, 483-485 (1996). Revised for second edition (2001).
Data-Parallel Implementations of Dense Simplex Methods on the Connection Machine CM-2: with I. Boduroglu, L. Polymenakos, and D. Goldfarb. ORSA Journal on Computing 7(4):402-416 (1995).
Parallel Computing in Network Optimization: with D. P. Bertsekas (lead author), D. Castanon, and S. A. Zenios. Network Models, M. O. Ball, T. L. Magnanti, C. L. Monma, and G. L. Nemhauser, eds., Handbooks in Operations Research and Management Science 7, Elsevier Scientific, Amsterdam, 331-399 (1995).
Control Strategies for Parallel Mixed Integer Branch and Bound: Proceedings of Supercomputing'94, IEEE Computer Society Press, Los Alamitos, CA, 41-48 (1994).
Parallel Branch-and-Bound Methods for Mixed-Integer Programming on the CM-5: SIAM Journal on Optimization 4(4):794-814 (1994). This article includes and supersedes Thinking Machines technical report TMC-257, which persists in many out-of-date reference lists.
Some Reformulations and Applications of the Alternating Direction Method of Multipliers: with M. Fukushima. In Large Scale Optimization: State of the Art, W. W. Hager, D. W. Hearn, P. M. Pardalos, eds., Kluwer Academic, Dordrecht, 115-134 (1994).
Parallel Branch-and-Bound Methods for Mixed Integer Programming: SIAM News 27(1):1,12-15 (1994). Available in revised form as Chapter 13 of the collection Applications on Advanced Architecture Computers, Greg Astfalk, editor, SIAM Books, 1996.
Some Saddle-Function Splitting Methods for Convex Programming: Optimization Methods and Software 4:75-83 (1994).
Alternating Direction Multiplier Decomposition of Convex Programs: Journal of Optimization Theory and Applications 80(1):39-62 (1994).
Stochastic Dedication: Designing Fixed-Income Portfolios using Massively Parallel Benders Decomposition: with R. S. Hiller (lead author). Management Science 39(11):1422-1438 (1993).
The Alternating Step Method for Monotropic Programming on the Connection Machine 2: ORSA Journal on Computing 5(1):84-96 (1993).
Nonlinear Proximal Point Algorithms using Bregman Functions, with Applications to Convex Programming: Mathematics of Operations Research 18(1):202-226 (1993).
Large-Scale Parallel Computing, Optimization, and Operations Research: A Survey: ORSA Computer Science Technical Section Newsletter 14(2), 1, 8-12, 25-28 (1993).
On the Douglas-Rachford Splitting Method and the Proximal Point Algorithm for Maximal Monotone Operators: with D. P. Bertsekas. Mathematical Programming 55(3):293-318 (1992).
Dual Coordinate Step Methods for Linear Network Flow Problems: with D. P. Bertsekas. Mathematical Programming 42(2):203-243 (1988).
Distributed Asynchronous Relaxation Methods for Linear Network Flow Problems: with D. P. Bertsekas. Proceedings of the International Federation of Automatic Control (IFAC), Munich (1987).
Optimization of Group Line-Haul Operations for Motor Carriers Using Twin Trailers: with Y. Sheffi. Transportation Research Record 1120:12-23 (1987).

Doctoral Dissertation

By popular demand, you can now download copies of my doctoral dissertation.

Scanned version (about 16 MB)
I hoped at one point to produce a more compact version from the orginal Macintosh Microsoft Word files, but it appears that is too difficult due to the nonstandard fonts and the difficulty of translating the formatting.

Miscellaneous

Lasso datasets (MATLAB format)

I am the parent of an autistic child (Ben). If you would like to join the fight against autism, I suggest supporting Autism Speaks or an intensive ABA/DTT/Verbal Behavior provider or consultant.

I play the tenor saxophone -- mostly jazz and Brazilian music.

Playing Aquarella do Brasil with Carlos Oliviera, March 2009
"Flagg Street", original composition, dedicated to the memory of Paul Tseng

I also play some jazz-like guitar in a much more amateurish way.

Slice of Life, composed and directed by my son Caleb

Here are my musical toys (instruments)...

Musical performances and sheet music dedicated to the memory of Paul Tseng.

The blog that ate New Jersey

The "vi" text editor is for troglodytes. I strongly preferred the GNU EMACS text editor. But eventually I was won over by Sublime Text. It's not free, but one license serves forever on any combination of platforms.

Back when mastodons roamed New Jersey, memory really was "core", and it was measured in K, I was a member of the R.E.S.I.S.T.O.R.S. computer club, and even president during some of its waning days.

My favorite movie is Black Orpheus. I know it is not a very accurate picture of Brazil, but it is incredible cinema with wonderful music.

I am a fan of silly/surreal British humor, especially Monty Python (I love "The Argument Clinic") and The Goon Show.

Pointers to some of my favorite music.

Friends and relatives on the web:

Schultz Writing Services
My late father, Harry Eckstein, who was one of the (approximately) one thousand children of Jewish descent who came to the US from Germany without their parents during the Nazi period.
My friend David Erle.
My friend Tom Levenson.
My friend Brian Pentland.
My friend Jeff Ritter.
My friend Jordan Sand.
My friend Bob Vanderbei.

"Homestar runner" -- accept no substitutes! (Caution: requires "flash" and contains (sort of) adult language)