Angelos Tsoukalas
Research Subjects
Angelos Tsoukalas has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Alexander Mitsos
RWTH Aachen University
7 shared publications - Andrew Eberhard
School of Mathematical Sciences, RMIT University, Melbourne, Australia
3 shared publications - Angelos Georghiou
Department of Business and Public Administration, University of Cyprus, Nicosia 1678, Cyprus
3 shared publications - Dominik Bongartz
Department of Chemical Engineering, KU Leuven
3 shared publications - Natashia Boland
School of Electrical Engineering, Computing and Mathematical Sciences Curtin University Perth Western Australia
3 shared publications - Susanne Sass
Process Systems Engineering (AVT.SVT), RWTH Aachen University, Aachen, Germany
3 shared publications
Latest Publications
- 1 February 2026
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Management Science
- Vol. 72 (2) , 1456-1471
On the Optimality of Affine Decision Rules in Distributionally Robust Optimization
We propose conditions under which two-stage distributionally robust optimization problems are optimally solved in affine or K-adaptable affine decision rules. Contrary to previous work, our conditions do not impose any structure on the support of the uncertain parameters, and they ensure pointwise (as opposed to worst case) optimality of (K-adaptable) affine decision rules. The absence of support restrictions allows us to transfer nonlinearities from the problem description to the support via liftings, whereas the pointwise optimality implies that decision rules remain optimal for broad classes of distributionally robust optimization problems, including data-driven problems over [ ]-divergence or Wasserstein ambiguity sets. We demonstrate how our conditions can be met in two applications. This paper was accepted by Chung Piaw Teo, optimization and decision analytics. Funding: This work was supported by Engineering and Physical Sciences Research Council (EPSRC) [Grant EP/W003317/1]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/mnsc.2023.00053 .
- 1 February 2026
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Management Science
- Vol. 72 (2) , 1456-1471
On the Optimality of Affine Decision Rules in Distributionally Robust Optimization
We propose conditions under which two-stage distributionally robust optimization problems are optimally solved in affine or K-adaptable affine decision rules. Contrary to previous work, our conditions do not impose any structure on the support of the uncertain parameters, and they ensure pointwise (as opposed to worst case) optimality of (K-adaptable) affine decision rules. The absence of support restrictions allows us to transfer nonlinearities from the problem description to the support via liftings, whereas the pointwise optimality implies that decision rules remain optimal for broad classes of distributionally robust optimization problems, including data-driven problems over [ ]-divergence or Wasserstein ambiguity sets. We demonstrate how our conditions can be met in two applications. This paper was accepted by Chung Piaw Teo, optimization and decision analytics. Funding: This work was supported by Engineering and Physical Sciences Research Council (EPSRC) [Grant EP/W003317/1]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/mnsc.2023.00053 .
- Open Access
- 23 June 2025
- Published by Springer Nature in Journal of Global Optimization
- Vol. 92 (3) , 615-642
Out-of-sample estimation for a branch-and-bound algorithm with growing datasets
In [Sass et al., Eur. J. Oper. Res., 316 (1): 36 – 45, 2024], we proposed a branch-and-bound (B&B) algorithm with growing datasets for the deterministic global optimization of parameter estimation problems based on large datasets. Therein, we start the B&B algorithm with a reduced dataset and augment it until reaching the full dataset upon convergence. However, convergence may be slowed down by a gap between the lower bounds of the reduced and the original problem, in particular for noisy measurement data. Thus, we propose the use of out-of-sample estimation for improving the lower bounds calculated with reduced datasets. Based on this, we extend the deterministic approach and propose two heuristic approaches. The computational performance of all approaches is compared with the standard B&B algorithm as a benchmark based on real-world estimation problems from process systems engineering, biochemistry, and machine learning covering datasets with and without measurement noise. Our results indicate that the heuristic approaches can improve the final lower bounds on the optimal objective value without cutting off the global solution. Aside from this, we prove that resampling can decrease the variance of the lower bounds calculated based on random initial datasets. In our case study, resampling hardly affects the performance of the approaches which indicates that the B&B algorithm with growing datasets does not suffer from large variances.
- 23 June 2025
- Published by Springer Nature in Journal of Global Optimization
- Vol. 92 (3) , 615-642
Open AccessOut-of-sample estimation for a branch-and-bound algorithm with growing datasets
In [Sass et al., Eur. J. Oper. Res., 316 (1): 36 – 45, 2024], we proposed a branch-and-bound (B&B) algorithm with growing datasets for the deterministic global optimization of parameter estimation problems based on large datasets. Therein, we start the B&B algorithm with a reduced dataset and augment it until reaching the full dataset upon convergence. However, convergence may be slowed down by a gap between the lower bounds of the reduced and the original problem, in particular for noisy measurement data. Thus, we propose the use of out-of-sample estimation for improving the lower bounds calculated with reduced datasets. Based on this, we extend the deterministic approach and propose two heuristic approaches. The computational performance of all approaches is compared with the standard B&B algorithm as a benchmark based on real-world estimation problems from process systems engineering, biochemistry, and machine learning covering datasets with and without measurement noise. Our results indicate that the heuristic approaches can improve the final lower bounds on the optimal objective value without cutting off the global solution. Aside from this, we prove that resampling can decrease the variance of the lower bounds calculated based on random initial datasets. In our case study, resampling hardly affects the performance of the approaches which indicates that the B&B algorithm with growing datasets does not suffer from large variances.
- Open Access
- 23 February 2024
- Published by Elsevier in European Journal of Operational Research
- Vol. 316 (1) , 36-45
A branch-and-bound algorithm with growing datasets for large-scale parameter estimation
- 23 February 2024
- Published by Elsevier in European Journal of Operational Research
- Vol. 316 (1) , 36-45
Open AccessA branch-and-bound algorithm with growing datasets for large-scale parameter estimation
- 1 July 2023
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 71 (4) , 1260-1276
Revenue Management of a Professional Services Firm with Quality Revelation
Professional service firms (PSFs) such as management consulting, law, accounting, investment banking, architecture, advertising, and home-repair companies provide services for complicated turnkey projects. A firm bids for a project and, if successful in the bid, assigns employees to work on the project. We formulate this as a revenue management problem under two assumptions: a quality-revelation setup, where the employees that would be assigned to the project are committed ex ante, as part of the bid, and a quality-reputation setup, where the bid’s win probability depends on past performance, say, an average of the quality of past jobs. We first model a stylized Markov chain model of the problem amenable to analysis and show that up-front revelation of the assigned employees has subtle advantages. Subsequent to this analysis, we develop an operational stochastic dynamic programming framework under the revelation model to aid the firm in this bidding and assignment process. We show that the problem is computationally challenging and provide a series of bounds and solution methods to approximate the stochastic dynamic program. Based on our model and computational methods, we are able to address a number of interesting business questions for a PSF, such as the optimal utilization levels and the value of each employee type. Our methodology provides management with a tool kit for bidding on projects as well as to perform workforce analytics and to make staffing decisions.
- 1 July 2023
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 71 (4) , 1260-1276
Revenue Management of a Professional Services Firm with Quality Revelation
Professional service firms (PSFs) such as management consulting, law, accounting, investment banking, architecture, advertising, and home-repair companies provide services for complicated turnkey projects. A firm bids for a project and, if successful in the bid, assigns employees to work on the project. We formulate this as a revenue management problem under two assumptions: a quality-revelation setup, where the employees that would be assigned to the project are committed ex ante, as part of the bid, and a quality-reputation setup, where the bid’s win probability depends on past performance, say, an average of the quality of past jobs. We first model a stylized Markov chain model of the problem amenable to analysis and show that up-front revelation of the assigned employees has subtle advantages. Subsequent to this analysis, we develop an operational stochastic dynamic programming framework under the revelation model to aid the firm in this bidding and assignment process. We show that the problem is computationally challenging and provide a series of bounds and solution methods to approximate the stochastic dynamic program. Based on our model and computational methods, we are able to address a number of interesting business questions for a PSF, such as the optimal utilization levels and the value of each employee type. Our methodology provides management with a tool kit for bidding on projects as well as to perform workforce analytics and to make staffing decisions.
- Open Access
- 9 May 2023
- Published by Taylor & Francis in Optimization Methods and Software
- Vol. 38 (6) , 1129-1141
Towards global parameter estimation exploiting reduced data sets
We focus on deterministic global optimization (DGO) for nonconvex parameter estimation problems. Realistic and accurate solutions often require a fitting against very large measurement data sets, resulting in intractable size for DGO. Thus, we aim at accelerating the branch-and-bound algorithm by using reduced data sets for constructing valid bounds. We focus on fitting the equation of state for propane, which is of high interest for the chemical industry. The resulting estimation problem is a challenging nonconvex mixed-integer nonlinear optimization problem. We investigate the validity of using reduced data sets by comparing how the lower and upper bounds change when replacing the full data set with different reduced data sets. We account for regions with high and low quality fits by considering the results for the whole feasible region and 100 different subregions. Our results indicate that both regions containing solution candidates and regions containing only low quality fits can be identified based on reduced data sets. Moreover, we observe that the average CPU time per branch-and-bound iteration typically decreases if reduced data sets are used.
- 9 May 2023
- Published by Taylor & Francis in Optimization Methods and Software
- Vol. 38 (6) , 1129-1141
Open AccessTowards global parameter estimation exploiting reduced data sets
We focus on deterministic global optimization (DGO) for nonconvex parameter estimation problems. Realistic and accurate solutions often require a fitting against very large measurement data sets, resulting in intractable size for DGO. Thus, we aim at accelerating the branch-and-bound algorithm by using reduced data sets for constructing valid bounds. We focus on fitting the equation of state for propane, which is of high interest for the chemical industry. The resulting estimation problem is a challenging nonconvex mixed-integer nonlinear optimization problem. We investigate the validity of using reduced data sets by comparing how the lower and upper bounds change when replacing the full data set with different reduced data sets. We account for regions with high and low quality fits by considering the results for the whole feasible region and 100 different subregions. Our results indicate that both regions containing solution candidates and regions containing only low quality fits can be identified based on reduced data sets. Moreover, we observe that the average CPU time per branch-and-bound iteration typically decreases if reduced data sets are used.
- 1 March 2020
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 68 (2) , 572-590
A Primal-Dual Lifting Scheme for Two-Stage Robust Optimization
Two-stage robust optimization problems, in which decisions are taken both in anticipation of and in response to the observation of an unknown parameter vector from within an uncertainty set, are notoriously challenging. In this paper, we develop convergent hierarchies of primal (conservative) and dual (progressive) bounds for these problems that trade off the competing goals of tractability and optimality: Although the coarsest bounds recover a tractable but suboptimal affine decision rule approximation of the two-stage robust optimization problem, the refined bounds lift extreme points of the uncertainty set until an exact but intractable extreme point reformulation of the problem is obtained. Based on these bounds, we propose a primal-dual lifting scheme for the solution of two-stage robust optimization problems that accommodates for discrete, here-and-now decisions, infeasible problem instances, and the absence of a relatively complete recourse. The incumbent solutions in each step of our algorithm afford rigorous error bounds, and they can be interpreted as piecewise affine decision rules. We illustrate the performance of our algorithm on illustrative examples and on an inventory management problem.
- 1 March 2020
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 68 (2) , 572-590
A Primal-Dual Lifting Scheme for Two-Stage Robust Optimization
Two-stage robust optimization problems, in which decisions are taken both in anticipation of and in response to the observation of an unknown parameter vector from within an uncertainty set, are notoriously challenging. In this paper, we develop convergent hierarchies of primal (conservative) and dual (progressive) bounds for these problems that trade off the competing goals of tractability and optimality: Although the coarsest bounds recover a tractable but suboptimal affine decision rule approximation of the two-stage robust optimization problem, the refined bounds lift extreme points of the uncertainty set until an exact but intractable extreme point reformulation of the problem is obtained. Based on these bounds, we propose a primal-dual lifting scheme for the solution of two-stage robust optimization problems that accommodates for discrete, here-and-now decisions, infeasible problem instances, and the absence of a relatively complete recourse. The incumbent solutions in each step of our algorithm afford rigorous error bounds, and they can be interpreted as piecewise affine decision rules. We illustrate the performance of our algorithm on illustrative examples and on an inventory management problem.
- 1 May 2019
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 67 (3) , 813-830
Robust Dual Dynamic Programming
In the paper “Robust Dual Dynamic Programming,” Angelos Georghiou, Angelos Tsoukalas, and Wolfram Wiesemann propose a novel solution scheme for addressing planning problems with long horizons. Such problems can be formulated as multistage robust optimization problems. The proposed method takes advantage of the decomposable nature of these problems by bounding the costs arising in the future stages through lower and upper cost-to-go functions. The proposed scheme does not require a relatively complete recourse, and it offers deterministic upper and lower bounds throughout the execution of the algorithm. The promising performance of the algorithm is shown in a stylized inventory-management problem in which the proposed algorithm achieved the optimal solution in problem instances with 100 time stages in a few minutes.
- 1 May 2019
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 67 (3) , 813-830
Robust Dual Dynamic Programming
In the paper “Robust Dual Dynamic Programming,” Angelos Georghiou, Angelos Tsoukalas, and Wolfram Wiesemann propose a novel solution scheme for addressing planning problems with long horizons. Such problems can be formulated as multistage robust optimization problems. The proposed method takes advantage of the decomposable nature of these problems by bounding the costs arising in the future stages through lower and upper cost-to-go functions. The proposed scheme does not require a relatively complete recourse, and it offers deterministic upper and lower bounds throughout the execution of the algorithm. The promising performance of the algorithm is shown in a stylized inventory-management problem in which the proposed algorithm achieved the optimal solution in problem instances with 100 time stages in a few minutes.
- 25 October 2016
- Published by Springer Nature in Journal of Global Optimization
- Vol. 68 (1) , 219-225
Erratum to: Multivariate McCormick relaxations
- 25 October 2016
- Published by Springer Nature in Journal of Global Optimization
- Vol. 68 (1) , 219-225
Erratum to: Multivariate McCormick relaxations
- 18 November 2014
- Published by Springer Nature in Journal of Optimization Theory and Applications
- Vol. 167 (2) , 558-584
A Trust Region Method for the Solution of the Surrogate Dual in Integer Programming
- 18 November 2014
- Published by Springer Nature in Journal of Optimization Theory and Applications
- Vol. 167 (2) , 558-584
A Trust Region Method for the Solution of the Surrogate Dual in Integer Programming
- 12 April 2014
- Published by Springer Nature in Mathematical Programming Computation
- Vol. 6 (3) , 255-279
Boosting the feasibility pump
- 12 April 2014
- Published by Springer Nature in Mathematical Programming Computation
- Vol. 6 (3) , 255-279
Boosting the feasibility pump