Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

IEEE Access

QUBO Formulations and Characterization of Penalty Parameters for the Multi-Knapsack Problem

Evren Güney, Joachim C.F. Ehrenthal, Thomas Hanne

Published Jan 1, 2025• 11 citations
Publisher

Abstract

The Multi-Knapsack Problem (MKP) is a fundamental challenge in operations research and combinatorial optimization. Quantum computing introduces new possibilities for solving MKP using Quadratic Unconstrained Binary Optimization (QUBO) models. However, a key challenge in QUBO formulations is the selection of penalty parameters, which directly influence solution feasibility and algorithm performance. In this work, we develop QUBO formulations for two MKP variants—the Multidimensional Knapsack Problem (MDKP) and the Multiple Knapsack Problem (MUKP)—and provide an algebraic characterization of their penalty parameters. We systematically evaluate their impact through quantum simulation experiments and compare the performance of the two leading quantum optimization approaches: Quantum Approximate Optimization Algorithm (QAOA) and quantum annealing, alongside a state-of-the-art classical solver. Our results indicate that while classical solvers remain superior, careful tuning of penalty parameters has a strong impact on quantum optimization outcomes. QAOA is highly sensitive to parameter choices, whereas quantum annealing produces more stable results on small to mid-sized instances. Further, our results reveal that MDKP instances can maintain feasibility at penalty values below theoretical bounds, while MUKP instances show greater sensitivity to penalty reductions. Finally, we outline directions for future research in solving MKP, including adaptive penalty parameter tuning, hybrid quantum-classical approaches, and practical optimization strategies for QAOA, as well as real-hardware evaluations.

DOI
10.1109/access.2025.3550788

Related papers

Quantum Approaches to the 0/1 Multi-Knapsack Problem: QUBO Formulation, Penalty Parameter Characterization and Analysis

Evren Güney, Joachim C.F. Ehrenthal, Thomas Hanne

The 0/1 Multi-Knapsack Problem (MKP) is a combinatorial optimization problem with applications in lo gistics, finance, and resource management. Advances in quantum computing have enabled the exploration of problems like the 0/1 MKP through Quadratic Unconstrained Binary Optimization (QUBO) formulations. This work develops QUBO formulations for the 0/1 MKP, with a focus on optimizing penalty parameters for encoding constraints. Using simulation experiments across quantum platforms, we evaluate the feasibility of solving small-scale instances of the 0/1 MKP. The results provide insights into the challenges and opportuni ties associated with applying quantum optimization methods for constrained resource allocation problems. © 2025 by SCITEPRESS– Science and Technology Publications, Lda.

International Conference on Agents and Artificial Intelligence (ICAART)•Jan 1, 2025•1 citation

Resource-efficient formulations of the Knapsack problem for quantum computing

Pim van den Heuvel, Vrinda Mehta, Dennis Willsch +2 more

Quantum computing has emerged as a new tool for solving combinatorial optimization problems. Quantum optimization approaches usually operate on the subclass of Quadratic Unconstrained Binary Optimization (QUBO) problems. Many optimization problems can be reformulated in this form, though there are often multiple ways to do so. The choice of formulation can considerably affect the performance of QUBO solvers in practice. In this work, we study QUBO formulations of the Knapsack problem, a canonical example of a problem with inequality constraints. In the standard formulation these constraints are converted into penalty terms, which requires additional slack variables. Here, we investigate three more resource-efficient approaches for formulating the Knapsack as a QUBO. The first two modify the standard formulation, either by deriving a smaller penalty constant or by reducing the number of slack variables. The third approach is a wholly different method, which avoids the use of slack variables entirely. All three approaches increase the probability of finding the optimal solution, both in simulations of the linear ramp Quantum Approximate Optimization Algorithm (LR-QAOA) and on a D-Wave quantum annealer. The slackless method performs best, increasing success probabilities by several orders of magnitude even for relatively small problem sizes.

arXiv:2611.06797•Nov 4, 2026•0 citations

QUBO Formulations and Quantum Optimization for the Multi-Dimensional Knapsack Problem With Conflict, Forcing, and Precedence Constraints

Evren Güney, Joachim C.F. Ehrenthal

Constrained knapsack variants are well-suited for QUBO-based quantum optimization, but adding logical relations can inflate the binary model and complicate penalty selection. This work examines the multi-dimensional knapsack problems augmented with conflict, forcing, and precedence (CFP) constraints using Quadratic Unconstrained Binary Optimization (QUBO) models across classical and quantum platforms. The additional logical constraints are incorporated through quadratic penalty terms that modify existing QUBO coefficients without introducing additional binary variables beyond the slack variables used for the capacity constraints. Computational studies are conducted to assess how each of the three constraint types affects solution behavior, including solution quality, computation time, and sensitivity to penalty parameter magnitudes. The approach is demonstrated on representative small-scale instances using QAOA on IBM Quantum gate-based device and quantum annealing on a D-Wave Advantage processor, with results compared with Gurobi as classical solver. The experiments show that the proposed QUBO formulations achieve optimal or near-optimal solutions for small instances, while showing the sensitivity of current quantum hardware to penalty selection, and scaling. Overall, CFP penalties are variable-neutral relative to the slack-based capacity representation, while the capacity terms dominate logical-qubit count and coupling density. Future work may target more compact or hardware-aware capacity modeling.

IEEE Access•Jan 1, 2026•2 citations

Resource-efficient QUBO formulations for the Knapsack and Bin Packing problem

Pim van den Heuvel

Quantum computing approaches, such as Quantum Annealing (QA) and the Quantum Approximate Optimization Algorithm (QAOA), have emerged as promising candidates for solving combinatorial optimization problems. Since these approaches typically work with Quadratic Unconstrained Binary Optimization (QUBO) problems, constraints, when present, are commonly encoded as penalty terms. Conventional approaches to incorporate inequality constraints into QUBO formulations additionally require the introduction of slack variables, increasing the required number of qubits. The Knapsack and Bin Packing problems are optimization problems with such inequality constraints. We show that the number of slack variables required to formulate the Knapsack problem as a QUBO problem can be reduced using certain properties specific to this problem. Moreover, we present two separate strategies for the Knapsack and Bin Packing problems that avoid using the costly slack variables altogether. Instead, both strategies solve several QUBO problems, with varying penalty constants, such that at least one of them is a good representation of the given optimization problem. Numerical simulations of QAOA, as well as experiments on a D-Wave quantum annealer, show that these slack-less approaches can improve performance compared to the conventional slack variable method (in terms of success probability), even when the latter is allowed to vary penalty constants too.

•Jan 1, 2025•0 citations

QUBO formulations for parallel machine scheduling

Frank Phillipson, Yoseph Y. Hutten

This paper investigates several Quadratic Unconstrained Binary optimisation (QUBO) formulations for the parallel machine scheduling problem P m | | C max with the goal of assessing their suitability for quantum annealing and related heuristic optimisation methods. Three QUBO models are constructed, ranging from formulations derived from classical integer and one-hot encodings to a more compact load-balancing based representation. Extensive computational experiments using Simulated Annealing (SA) and Quantum Annealing (QA) reveal substantial differences in scalability and solution quality among the formulations. The alternative load-balanced model consistently yields near-optimal makespans, typically within 1%–3% of the theoretical lower bound, highlighting its stability and effectiveness compared to the more penalty sensitive formulations, whose performance deteriorates for larger instances. Sensitivity analyses further reveal that the appropriate selection of penalty weights and problem-specific parameter tuning significantly influences both solution accuracy and computational effort in QUBO formulations. While current quantum hardware limitations restrict experimentation to small problem sizes, the findings illustrate how compact QUBO designs can better exploit quantum optimisation frameworks. The insights gained in this study provide practical modelling guidelines for future quantum scheduling research and lay the groundwork for extending QUBO-based approaches to more complex scheduling environments. • Three QUBO formulations are compared for identical parallel machine scheduling. • A compact load-balancing QUBO avoids auxiliary variables. • Load balancing yields makespans on average within 1–3%. • Both simulated and quantum annealing produce near-optimal schedules. • Penalty selection and encoding strongly affect scalability and solution quality.

Computers & Operations Research•Sep 11, 2026•0 citations

A Quantum Search Method for Quadratic and Multidimensional Knapsack Problems

Sören Wilkening, Andreea-Iulia Lefterovici, Lennart Binkowski +5 more

Solving combinatorial optimization problems is a promising application area for quantum algorithms in real-world scenarios. In this work, we extend the “Quantum Tree Generator” (QTG), previously proposed for the 0-1 Knapsack Problem, to the 0-1 Quadratic Knapsack Problem (QKP) and the Multidimensional Knapsack Problem (MDKP). The QTG constructs a superposition of all feasible solutions for a given instance and can therefore be utilized as a promising state preparation routine within amplitude amplification to produce high-quality solutions. Previously, QTG-based search was tested on the 0-1 Knapsack Problem, where it demonstrated the potential for practical quantum advantage, once quantum computers with a few hundred logical and fully connected qubits are available. Here, we evaluate the algorithm's performance on QKP and MDKP against the classical solver Gurobi. To facilitate large-scale evaluations, we employ an advanced benchmarking technique that enables runtime predictions for instances with up to 2000 variables for QKP and up to 1500 variables and 100 constraints for MDKP. Our results indicate that QTG-based search can produce high-quality solutions with competitive runtimes for QKP. However, its performance declines for MDKP, highlighting the challenges quantum algorithms face when tackling highly constrained optimization problems.

arXiv:2503.22325•IEEE International Conference on Quantum Computing and Engineering (QCE)•Aug 30, 2025•3 citations
The Multi-Knapsack Problem (MKP) is a fundamental challenge in operations research and combinatorial optimization. Quantum computing introduces new possibilities for solving MKP using Quadratic Unconstrained Binary Optimization (QUBO) models. However, a key challenge in QUBO formulations is the selection of penalty parameters, which directly influence solution feasibility and algorithm performance. In this work, we develop QUBO formulations for two MKP variants—the Multidimensional Knapsack Problem (MDKP) and the Multiple Knapsack Problem (MUKP)—and provide an algebraic characterization of their penalty parameters. We systematically evaluate their impact through quantum simulation experiments and compare the performance of the two leading quantum optimization approaches: Quantum Approximate Optimization Algorithm (QAOA) and quantum annealing, alongside a state-of-the-art classical solver. Our results indicate that while classical solvers remain superior, careful tuning of penalty parameters has a strong impact on quantum optimization outcomes. QAOA is highly sensitive to parameter choices, whereas quantum annealing produces more stable results on small to mid-sized instances. Further, our results reveal that MDKP instances can maintain feasibility at penalty values below theoretical bounds, while MUKP instances show greater sensitivity to penalty reductions. Finally, we outline directions for future research in solving MKP, including adaptive penalty parameter tuning, hybrid quantum-classical approaches, and practical optimization strategies for QAOA, as well as real-hardware evaluations.
The 0/1 Multi-Knapsack Problem (MKP) is a combinatorial optimization problem with applications in lo gistics, finance, and resource management. Advances in quantum computing have enabled the exploration of problems like the 0/1 MKP through Quadratic Unconstrained Binary Optimization (QUBO) formulations. This work develops QUBO formulations for the 0/1 MKP, with a focus on optimizing penalty parameters for encoding constraints. Using simulation experiments across quantum platforms, we evaluate the feasibility of solving small-scale instances of the 0/1 MKP. The results provide insights into the challenges and opportuni ties associated with applying quantum optimization methods for constrained resource allocation problems. © 2025 by SCITEPRESS– Science and Technology Publications, Lda.
Quantum computing has emerged as a new tool for solving combinatorial optimization problems. Quantum optimization approaches usually operate on the subclass of Quadratic Unconstrained Binary Optimization (QUBO) problems. Many optimization problems can be reformulated in this form, though there are often multiple ways to do so. The choice of formulation can considerably affect the performance of QUBO solvers in practice. In this work, we study QUBO formulations of the Knapsack problem, a canonical example of a problem with inequality constraints. In the standard formulation these constraints are converted into penalty terms, which requires additional slack variables. Here, we investigate three more resource-efficient approaches for formulating the Knapsack as a QUBO. The first two modify the standard formulation, either by deriving a smaller penalty constant or by reducing the number of slack variables. The third approach is a wholly different method, which avoids the use of slack variables entirely. All three approaches increase the probability of finding the optimal solution, both in simulations of the linear ramp Quantum Approximate Optimization Algorithm (LR-QAOA) and on a D-Wave quantum annealer. The slackless method performs best, increasing success probabilities by several orders of magnitude even for relatively small problem sizes.
Constrained knapsack variants are well-suited for QUBO-based quantum optimization, but adding logical relations can inflate the binary model and complicate penalty selection. This work examines the multi-dimensional knapsack problems augmented with conflict, forcing, and precedence (CFP) constraints using Quadratic Unconstrained Binary Optimization (QUBO) models across classical and quantum platforms. The additional logical constraints are incorporated through quadratic penalty terms that modify existing QUBO coefficients without introducing additional binary variables beyond the slack variables used for the capacity constraints. Computational studies are conducted to assess how each of the three constraint types affects solution behavior, including solution quality, computation time, and sensitivity to penalty parameter magnitudes. The approach is demonstrated on representative small-scale instances using QAOA on IBM Quantum gate-based device and quantum annealing on a D-Wave Advantage processor, with results compared with Gurobi as classical solver. The experiments show that the proposed QUBO formulations achieve optimal or near-optimal solutions for small instances, while showing the sensitivity of current quantum hardware to penalty selection, and scaling. Overall, CFP penalties are variable-neutral relative to the slack-based capacity representation, while the capacity terms dominate logical-qubit count and coupling density. Future work may target more compact or hardware-aware capacity modeling.
Quantum computing approaches, such as Quantum Annealing (QA) and the Quantum Approximate Optimization Algorithm (QAOA), have emerged as promising candidates for solving combinatorial optimization problems. Since these approaches typically work with Quadratic Unconstrained Binary Optimization (QUBO) problems, constraints, when present, are commonly encoded as penalty terms. Conventional approaches to incorporate inequality constraints into QUBO formulations additionally require the introduction of slack variables, increasing the required number of qubits. The Knapsack and Bin Packing problems are optimization problems with such inequality constraints. We show that the number of slack variables required to formulate the Knapsack problem as a QUBO problem can be reduced using certain properties specific to this problem. Moreover, we present two separate strategies for the Knapsack and Bin Packing problems that avoid using the costly slack variables altogether. Instead, both strategies solve several QUBO problems, with varying penalty constants, such that at least one of them is a good representation of the given optimization problem. Numerical simulations of QAOA, as well as experiments on a D-Wave quantum annealer, show that these slack-less approaches can improve performance compared to the conventional slack variable method (in terms of success probability), even when the latter is allowed to vary penalty constants too.
This paper investigates several Quadratic Unconstrained Binary optimisation (QUBO) formulations for the parallel machine scheduling problem P m | | C max with the goal of assessing their suitability for quantum annealing and related heuristic optimisation methods. Three QUBO models are constructed, ranging from formulations derived from classical integer and one-hot encodings to a more compact load-balancing based representation. Extensive computational experiments using Simulated Annealing (SA) and Quantum Annealing (QA) reveal substantial differences in scalability and solution quality among the formulations. The alternative load-balanced model consistently yields near-optimal makespans, typically within 1%–3% of the theoretical lower bound, highlighting its stability and effectiveness compared to the more penalty sensitive formulations, whose performance deteriorates for larger instances. Sensitivity analyses further reveal that the appropriate selection of penalty weights and problem-specific parameter tuning significantly influences both solution accuracy and computational effort in QUBO formulations. While current quantum hardware limitations restrict experimentation to small problem sizes, the findings illustrate how compact QUBO designs can better exploit quantum optimisation frameworks. The insights gained in this study provide practical modelling guidelines for future quantum scheduling research and lay the groundwork for extending QUBO-based approaches to more complex scheduling environments. • Three QUBO formulations are compared for identical parallel machine scheduling. • A compact load-balancing QUBO avoids auxiliary variables. • Load balancing yields makespans on average within 1–3%. • Both simulated and quantum annealing produce near-optimal schedules. • Penalty selection and encoding strongly affect scalability and solution quality.
Solving combinatorial optimization problems is a promising application area for quantum algorithms in real-world scenarios. In this work, we extend the “Quantum Tree Generator” (QTG), previously proposed for the 0-1 Knapsack Problem, to the 0-1 Quadratic Knapsack Problem (QKP) and the Multidimensional Knapsack Problem (MDKP). The QTG constructs a superposition of all feasible solutions for a given instance and can therefore be utilized as a promising state preparation routine within amplitude amplification to produce high-quality solutions. Previously, QTG-based search was tested on the 0-1 Knapsack Problem, where it demonstrated the potential for practical quantum advantage, once quantum computers with a few hundred logical and fully connected qubits are available. Here, we evaluate the algorithm's performance on QKP and MDKP against the classical solver Gurobi. To facilitate large-scale evaluations, we employ an advanced benchmarking technique that enables runtime predictions for instances with up to 2000 variables for QKP and up to 1500 variables and 100 constraints for MDKP. Our results indicate that QTG-based search can produce high-quality solutions with competitive runtimes for QKP. However, its performance declines for MDKP, highlighting the challenges quantum algorithms face when tackling highly constrained optimization problems.