Francesca Maggioni
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Francesca Maggioni has a total of 96 co-authors. Here are the top co-authors by number of shared publications:
- Elisabetta Allevi
Department of Economics and Management, University of Brescia, Brescia, Italy
7 shared publications - Renzo L. Ricca
University of Milano-Bicocca
7 shared publications - Andrea Spinelli
University of Bergamo
6 shared publications Marida Bertocchi
University of Bergamo
6 shared publications- Luca Bertazzi
Department of Economics and Management, University of Brescia, 25121 Brescia, Italy
5 shared publications - Giorgio Consigli
Department of Mathematics, Khalifa University of Science and Technology, Abu Dhabi, UAE
4 shared publications
Latest Publications
- Top Cited PapersOpen Access
- 8 January 2026
- Published by Springer Nature in Annals of Operations Research
- p. 1-42
Asset liability management under sequential stochastic dominance constraints
We consider a financial intermediary managing assets and liabilities exposed to several risk sources and seeking an optimal portfolio strategy to minimise the initial capital invested and the total risk associated with investment losses and financial debt. We formulate the problem as a multistage stochastic programming (MSP) model, with a time-consistent dynamic risk measure in the objective function to control the investment risk. To ensure that the intermediary’s financial equilibrium is preserved, we introduce a funding constraint in the model by enforcing in a time-consistent manner a sequential second-order stochastic dominance (SSD) of the portfolio return distribution over the liability distribution. We demonstrate that imposing the SSD constraint at the last-but-one stage is sufficient to enforce the SSD ordering at each stage. To deal with the computational burden of associated MSP, we develop a novel decomposition scheme integrating, for the first time in the literature, time-consistent dynamic risk measures and sequential stochastic dominance constraints. The proposed methodology is computationally validated on a case study developed on a property and casualty asset-liability management problem.
- 8 January 2026
- Published by Springer Nature in Annals of Operations Research
- p. 1-42
Top Cited PapersOpen AccessAsset liability management under sequential stochastic dominance constraints
We consider a financial intermediary managing assets and liabilities exposed to several risk sources and seeking an optimal portfolio strategy to minimise the initial capital invested and the total risk associated with investment losses and financial debt. We formulate the problem as a multistage stochastic programming (MSP) model, with a time-consistent dynamic risk measure in the objective function to control the investment risk. To ensure that the intermediary’s financial equilibrium is preserved, we introduce a funding constraint in the model by enforcing in a time-consistent manner a sequential second-order stochastic dominance (SSD) of the portfolio return distribution over the liability distribution. We demonstrate that imposing the SSD constraint at the last-but-one stage is sufficient to enforce the SSD ordering at each stage. To deal with the computational burden of associated MSP, we develop a novel decomposition scheme integrating, for the first time in the literature, time-consistent dynamic risk measures and sequential stochastic dominance constraints. The proposed methodology is computationally validated on a case study developed on a property and casualty asset-liability management problem.
- Open Access
Stochastic Fleet Size and Mix Consistent Vehicle Routing Problem for Last Mile Delivery
In this paper, we address the joint optimization of fleet size and mix, along with vehicle routing, under uncertain customer demand. We propose a two-stage stochastic mixed-integer programming model, where first-stage decisions concern the composition of the delivery fleet and the design of consistent baseline routes. In the second stage, approximate recourse actions are introduced to adapt the initial routes in response to realized customer demands. The objective is to minimize the total delivery cost, including vehicle acquisition, travel distance, and penalty costs for unserved demand. To tackle the computational challenges arising in realistic problem instances, we develop a path-based reformulation of the model and design a Kernel Search-based heuristic to enhance scalability. Computational experiments on small synthetic instances, generated through a population-density-based sampling approach, are conducted to validate the formulation and assess the effects of demand stochasticity through standard stochastic measures, after applying a scenario reduction technique. Additional tests on large-scale real-world instances, based on data from the Italian postal company, demonstrate the effectiveness of the proposed approach and provide managerial and practical insights.
Open AccessStochastic Fleet Size and Mix Consistent Vehicle Routing Problem for Last Mile Delivery
In this paper, we address the joint optimization of fleet size and mix, along with vehicle routing, under uncertain customer demand. We propose a two-stage stochastic mixed-integer programming model, where first-stage decisions concern the composition of the delivery fleet and the design of consistent baseline routes. In the second stage, approximate recourse actions are introduced to adapt the initial routes in response to realized customer demands. The objective is to minimize the total delivery cost, including vehicle acquisition, travel distance, and penalty costs for unserved demand. To tackle the computational challenges arising in realistic problem instances, we develop a path-based reformulation of the model and design a Kernel Search-based heuristic to enhance scalability. Computational experiments on small synthetic instances, generated through a population-density-based sampling approach, are conducted to validate the formulation and assess the effects of demand stochasticity through standard stochastic measures, after applying a scenario reduction technique. Additional tests on large-scale real-world instances, based on data from the Italian postal company, demonstrate the effectiveness of the proposed approach and provide managerial and practical insights.
- Open Access
- 3 December 2025
- Published by Elsevier in Transportation Research Part C: Emerging Technologies
- Vol. 183
A stochastic electric vehicle routing problem under uncertain energy consumption
- 3 December 2025
- Published by Elsevier in Transportation Research Part C: Emerging Technologies
- Vol. 183
Open AccessA stochastic electric vehicle routing problem under uncertain energy consumption
- Open Access
Neural networks for multi-horizon stochastic programming
This paper proposes a machine-learning-based solution approach for solving multi-horizon stochastic programs. The approach embeds a deep learning neural network into a multi-horizon stochastic program to approximate the recourse operational objective function. The proposed approach is demonstrated on a UK power system planning problem with uncertainty at investment and operational timescales. The results show that (1) the surrogate neural network performs well across three different architectures, (2) the proposed approach is up to 34.72 times faster than the direct solution of the monolithic deterministic equivalent counterpart, (3) the surrogate-based solutions yield comparable in-sample stability and improved out-of-sample performance relative to the deterministic equivalent, indicating better generalisation to unseen scenarios. The main contributions of the paper are: (1) we propose a machine-learning-based framework for solving multi-horizon stochastic programs, (2) we introduce a neural network embedding formulation tailored to multi-horizon stochastic programs with continuous first-stage decisions and fixed scenario sets, extending existing surrogate modelling approaches from two-stage to multi-horizon settings, and (3) we provide an extensive computational study on a realistic UK power system planning problem, demonstrating the trade-off between approximation accuracy, computational efficiency, and solution robustness for different neural network architectures and scenario set sizes.
Open AccessNeural networks for multi-horizon stochastic programming
This paper proposes a machine-learning-based solution approach for solving multi-horizon stochastic programs. The approach embeds a deep learning neural network into a multi-horizon stochastic program to approximate the recourse operational objective function. The proposed approach is demonstrated on a UK power system planning problem with uncertainty at investment and operational timescales. The results show that (1) the surrogate neural network performs well across three different architectures, (2) the proposed approach is up to 34.72 times faster than the direct solution of the monolithic deterministic equivalent counterpart, (3) the surrogate-based solutions yield comparable in-sample stability and improved out-of-sample performance relative to the deterministic equivalent, indicating better generalisation to unseen scenarios. The main contributions of the paper are: (1) we propose a machine-learning-based framework for solving multi-horizon stochastic programs, (2) we introduce a neural network embedding formulation tailored to multi-horizon stochastic programs with continuous first-stage decisions and fixed scenario sets, extending existing surrogate modelling approaches from two-stage to multi-horizon settings, and (3) we provide an extensive computational study on a realistic UK power system planning problem, demonstrating the trade-off between approximation accuracy, computational efficiency, and solution robustness for different neural network architectures and scenario set sizes.
- Open Access
- 11 September 2025
- Published by Elsevier in EURO Journal on Computational Optimization
- Vol. 13
A robust twin parametric margin support vector machine for multiclass classification
- 11 September 2025
- Published by Elsevier in EURO Journal on Computational Optimization
- Vol. 13
Open AccessA robust twin parametric margin support vector machine for multiclass classification
- Open Access
A Stochastic Electric Vehicle Routing Problem under Uncertain Energy Consumption
The increasing adoption of Electric Vehicles (EVs) for service and goods distribution operations has led to the emergence of Electric Vehicle Routing Problems (EVRPs), a class of vehicle routing problems addressing the unique challenges posed by the limited driving range and recharging needs of EVs. While the majority of EVRP variants have considered deterministic energy consumption, this paper focuses on the Stochastic Electric Vehicle Routing Problem with a Threshold recourse policy (SEVRP-T), where the uncertainty in energy consumption is considered, and a recourse policy is employed to ensure that EVs recharge at Charging Stations (CSs) whenever their State of Charge (SoC) falls below a specified threshold. We formulate the SEVRP-T as a two-stage stochastic mixed-integer second-order cone model, where the first stage determines the sequences of customers to be visited, and the second stage incorporates charging activities. The objective is to minimize the expected total duration of the routes, composed by travel times and recharging operations. To cope with the computational complexity of the model, we propose a heuristic based on an Iterated Local Search (ILS) procedure coupled with a Set Partitioning problem. To further speed up the heuristic, we develop two lower bounds on the corresponding first-stage customer sequences. Furthermore, to handle a large number of energy consumption scenarios, we employ a scenario reduction technique. Extensive computational experiments are conducted to validate the effectiveness of the proposed solution strategy and to assess the importance of considering the stochastic nature of the energy consumption. The research presented in this paper contributes to the growing body of literature on EVRP and provides insights into managing the operational deployment of EVs in logistics activities under uncertainty.
Open AccessA Stochastic Electric Vehicle Routing Problem under Uncertain Energy Consumption
The increasing adoption of Electric Vehicles (EVs) for service and goods distribution operations has led to the emergence of Electric Vehicle Routing Problems (EVRPs), a class of vehicle routing problems addressing the unique challenges posed by the limited driving range and recharging needs of EVs. While the majority of EVRP variants have considered deterministic energy consumption, this paper focuses on the Stochastic Electric Vehicle Routing Problem with a Threshold recourse policy (SEVRP-T), where the uncertainty in energy consumption is considered, and a recourse policy is employed to ensure that EVs recharge at Charging Stations (CSs) whenever their State of Charge (SoC) falls below a specified threshold. We formulate the SEVRP-T as a two-stage stochastic mixed-integer second-order cone model, where the first stage determines the sequences of customers to be visited, and the second stage incorporates charging activities. The objective is to minimize the expected total duration of the routes, composed by travel times and recharging operations. To cope with the computational complexity of the model, we propose a heuristic based on an Iterated Local Search (ILS) procedure coupled with a Set Partitioning problem. To further speed up the heuristic, we develop two lower bounds on the corresponding first-stage customer sequences. Furthermore, to handle a large number of energy consumption scenarios, we employ a scenario reduction technique. Extensive computational experiments are conducted to validate the effectiveness of the proposed solution strategy and to assess the importance of considering the stochastic nature of the energy consumption. The research presented in this paper contributes to the growing body of literature on EVRP and provides insights into managing the operational deployment of EVs in logistics activities under uncertainty.
- Open Access
- 18 June 2025
- Published by Elsevier in Decision Analytics Journal
- Vol. 16
A robust support vector machine approach for Raman data classification
- 18 June 2025
- Published by Elsevier in Decision Analytics Journal
- Vol. 16
Open AccessA robust support vector machine approach for Raman data classification
- Open Access
Asset liability management under sequential stochastic dominance constraints
We consider a financial intermediary managing assets and liabilities exposed to several risk sources and seeking an optimal portfolio strategy to minimise the initial capital invested and the total risk associated with investment losses and financial debt. We formulate the problem as a multistage stochastic programming model, with a time-consistent dynamic risk measure in the objective function to control the investment risk. To ensure that the intermediary's financial equilibrium is preserved, we introduce a funding constraint in the model by enforcing in a time-consistent manner a sequential second-order stochastic dominance (SSD) of the portfolio return distribution over the liability distribution. We demonstrate that imposing the SSD constraint at the last-but-one stage is sufficient to enforce the SSD ordering at each stage. To deal with the computational burden of associated MSP, we develop a novel decomposition scheme integrating, for the first time in the literature, time-consistent dynamic risk measures and sequential stochastic dominance constraints. The proposed methodology is computationally validated on a case study developed on a property and casualty ALM problem.
Open AccessAsset liability management under sequential stochastic dominance constraints
We consider a financial intermediary managing assets and liabilities exposed to several risk sources and seeking an optimal portfolio strategy to minimise the initial capital invested and the total risk associated with investment losses and financial debt. We formulate the problem as a multistage stochastic programming model, with a time-consistent dynamic risk measure in the objective function to control the investment risk. To ensure that the intermediary's financial equilibrium is preserved, we introduce a funding constraint in the model by enforcing in a time-consistent manner a sequential second-order stochastic dominance (SSD) of the portfolio return distribution over the liability distribution. We demonstrate that imposing the SSD constraint at the last-but-one stage is sufficient to enforce the SSD ordering at each stage. To deal with the computational burden of associated MSP, we develop a novel decomposition scheme integrating, for the first time in the literature, time-consistent dynamic risk measures and sequential stochastic dominance constraints. The proposed methodology is computationally validated on a case study developed on a property and casualty ALM problem.
- Open Access
- 29 April 2025
- Published by Elsevier in Computers & Operations Research
- Vol. 182
Robust and Distributionally Robust Shortest Path problems: A survey
- 29 April 2025
- Published by Elsevier in Computers & Operations Research
- Vol. 182
Open AccessRobust and Distributionally Robust Shortest Path problems: A survey
- Open Access
- 19 March 2025
- Published by Springer Nature in Annals of Operations Research
- Vol. 347 (3) , 1385-1423
Sampling methods for multi-stage robust optimization problems
In this paper, we consider multi-stage robust optimization problems of the minimax type. We assume that the total uncertainty set is the cartesian product of stagewise compact uncertainty sets and approximate the given problem by a sampled subproblem. Instead of looking for the worst case among the infinite and typically uncountable set of uncertain parameters, we consider only the worst case among a randomly selected subset of parameters. By adopting such a strategy, two main questions arise: (1) Can we quantify the error committed by the random approximation, especially as a function of the sample size? (2) If the sample size tends to infinity, does the optimal value converge to the “true” optimal value? Both questions will be answered in this paper. An explicit bound on the probability of violation is given and chain of lower bounds on the original multi-stage robust optimization problem provided. Numerical results dealing with a multi-stage inventory management problem show that the proposed approach works well for problems with two or three time periods while for larger ones the number of required samples is prohibitively large for computational tractability. Despite this, we believe that our results can be useful for problems with such small number of time periods, and it sheds some light on the challenge for problems with more time periods.
- 19 March 2025
- Published by Springer Nature in Annals of Operations Research
- Vol. 347 (3) , 1385-1423
Open AccessSampling methods for multi-stage robust optimization problems
In this paper, we consider multi-stage robust optimization problems of the minimax type. We assume that the total uncertainty set is the cartesian product of stagewise compact uncertainty sets and approximate the given problem by a sampled subproblem. Instead of looking for the worst case among the infinite and typically uncountable set of uncertain parameters, we consider only the worst case among a randomly selected subset of parameters. By adopting such a strategy, two main questions arise: (1) Can we quantify the error committed by the random approximation, especially as a function of the sample size? (2) If the sample size tends to infinity, does the optimal value converge to the “true” optimal value? Both questions will be answered in this paper. An explicit bound on the probability of violation is given and chain of lower bounds on the original multi-stage robust optimization problem provided. Numerical results dealing with a multi-stage inventory management problem show that the proposed approach works well for problems with two or three time periods while for larger ones the number of required samples is prohibitively large for computational tractability. Despite this, we believe that our results can be useful for problems with such small number of time periods, and it sheds some light on the challenge for problems with more time periods.