Enhancing Variational Quantum Algorithms for Multicriteria Optimization
Ivica Turkalj, Tom Ewen, Pascal Halffmann, Janik Maciejewski, Michael Trebing, Zakaria Abdelmoiz Dahi · 2025
Read Enhancing Variational Quantum Algorithms for Multicriteria Optimization by Ivica Turkalj, Tom Ewen, Pascal Halffmann, Janik Maciejewski, Michael Trebing, Zakaria Abdelmoiz Dahi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-22159.txt.
Introduction
Ivica Turkalj and co-authors wrote a paper in two thousand twenty-five on quantum methods for tasks with more than one goal. Such tasks require trade-offs, not a single best score. The authors work with variational quantum algorithms, which mix quantum circuits with classic computer search.
First, they recast the tuning of the circuit as a task with more than one goal, so classic search tools can be used at once. That mix beat the single-goal form on test tasks. Second, they add measures of spread to the cost score, which lets users control how wide the set of trade-off answers is. Spread rose by up to forty percent.
Catalog record
| CREATOR | Ivica Turkalj, Tom Ewen, Pascal Halffmann, Janik Maciejewski, Michael Trebing, Zakaria Abdelmoiz Dahi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 64 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-22159
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Enhancing Variational Quantum Algorithms for Multicriteria Optimization Thanks: This research was funded by Bundesministerium für Bildung und Forschung (BMBF), grant number 13N15695. Ivica Turkalj, Tom Ewen, Pascal Halffmann, Janik Maciejewski, Michael Trebing, Zakaria Abdelmoiz Dahi Abstract . This paper presents methodological improvements to variational quantum algorithms (VQAs) for solving multicriteria optimization problems. We introduce two key contributions. First, we reformulate the parameter optimization task of VQAs as a multicriteria problem, enabling the direct use of classical algorithms from various multicriteria metaheuristics. This hybrid framework outperforms the corresponding single-criteria VQAs in both average and worst-case performance across diverse benchmark problems. Second, we propose a method that augments the hypervolume-based cost function with coverage-oriented indicators, allowing explicit control over the diversity of the resulting Pareto front approximations. Experimental results show that our method can improve coverage by up to 40% with minimal loss in hypervolume. Our findings highlight the potential of combining quantum variational methods with classical population-based search to advance practical quantum optimization. 1. Introduction Quantum computing, by leveraging quantum mechanical phenomena such as superposition and entanglement, has demonstrated significant promise in solving complex computational problems. In particular, complex optimization problems, such as those labeled as NP-hard have been at the forefront at potential application areas, where quantum computing can solve these problems more efficient than classical resolution strategies [ AYCH23 ] . Among various quantum computing paradigms such as quantum annealing, variational quantum algorithms (VQAs) have emerged as a particularly compelling approach due to their hybrid quantum-classical nature, allowing current quantum devices, often limited by noise and coherence times, to solve meaningful problems. These algorithms iteratively adjust the parameters of quantum circuits (ansatz) to minimize or maximize objective functions, which are evaluated through quantum computations and optimized classically. Several studies observe advantages such as a better scaling, that is a better performance for large instance sizes [ SLC+24 ] , or better approximation quality [ PUW+24 ] . However, a true performance advantage in practice, either in solution time or quality has not been shown yet. Indeed, some publications hint that the limitations of current quantum optimization algorithms as QAOA prohibited any quantum advantage [ MSWB25 ] and claims of quantum supremacy of quantum annealing have quickly been debunked [ MC25 ] . A particular reason is the vast performance of current optimization solvers such as Gurobi that concentrate over 60 years of expertise in solving optimization problems in one solver framework. Therefore, benefits from applying quantum algorithms may arise first for complex optimization problems, for which high-performance classical solvers are lacking. Multicriteria optimization problems that deal with optimizing several, often conflicting objective function simultaneously are one of the most complex optimization problems. Besides their NP-hardness, there does not one optimal objective function value but several compromise solutions, so-called efficient or Pareto-optimal solutions with nondominated objective functions value exists. Finding all these compromises is computational demanding and even intractable due to the exponential (w.r.t. the encoding length of the problem) number of these compromises. While there exist many classical algorithms to tackle these problems, both exact methods [ HSD+22 ] and heuristics such as the famous Non-dominated Sorting Genetic Algorithm II (NSGAII) [ DPAM02 ] that is incorporated in the pymoo Python package [ BD20b ] , these algorithms are by far not as performant as solvers for e.g., mixed-integer linear problems. Since multicriteria optimization problems are omnipresent in practice, there is high need for faster and better multicriteria optimization algorithms and a chance for quantum algorithms to make a substantial impact, as it has been stated in the seminal overview on quantum optimization by Abbas et al. [ AAA+24 ] . 1.1. Recent Research There already exist several publications in the overlap of quantum computing and multicriteria optimization. Most of these publications are focused on the development of quantum-inspired algorithms to tackle multicriteria optimization problems, in particular heuristics such as evolutionary algorithms that take inspiration from quantum effects [ KKH ] . As these algorithms are solely classical approaches, we omit these in this overview. Multicriteria optimization has also been identified as a tool to improve quantum algorithms and their execution on real hardware. In particular, compiling a quantum circuit to the actual hardware while respecting the hardware architecture requirements is a multicriteria problem itself, e.g. one wants to minimize the circuit depth while minimize the number of ancilla qubits or error rates. Several papers have identified this multicriteria problem and provided multicriteria heuristics to improve the compiling, e.g. [ BCV21 , RB17 , SESS24 ] . Further, the paper [ DVLHHS+23 ] used a multicriteria view on the optimization of the parameters of variational algorithms for the resolution of optimization problems. They utilize NSGA-II to optimize the variational parameters subject to two objectives: the first one is the original objective of the optimization problem. The second objective is a function that evaluates the feasibility of the solution, i.e., it comprises of the penalty functions deduced from the constraints of the optimization problem. In a recent years, quantum methods also have been developed to directly target multicriteria problems. The first method has been published in 2016 [ BV16 ] and transforms the multicriteria problem into a single-criterion problem via weighted sum and solve this problem using D-Wave’s quantum annealer. Similar approaches using either QAOA or annealing have been presented for specific problems in [ CPM+24 ] for network routing and for portfolio optimization [ AdJP+24 , TAB+24 ] . Similarly, in [ UPMGRM22 ] the multicriteria problem is transformed into a single-criterion problem, the so-called lexicographic problem, where one objective is optimized and then added as constraint when optimizing the second objective function. The resulting single-criterion problems are solved using QAOA. In [ BV18 ] , the authors present a variant of Grover adaptive search, where the oracle flags solutions whose objective function values are not dominated. Given a starting point, this approach then generates an efficient solution. So far, all these approaches only provide a single efficient solution to the multicriteria problem. The first algorithms that aims at finding the whole Pareto front is presented in [ EWS24 ] . The authors propose a variant of QAOA for qudits instead of qubits, where for each objective function a problem Hamiltonian and a mixer Hamiltonian are incorporated in one layer of the QAOA. The variational parameters are then optimized with respect to the hypervolume of the obtained shots from the QAOA circuit execution. We will investigate this apporach in more detail in Section 3 . The authors in [ DCL+24 ] propose a decomposition-based Multi-Objective QAOA (MO-QAOA) using weighted sum and Tchebycheff scalarisation, while accounting for the constraints of current quantum hardware and simulators. Most recently, [ KPR+25 ] revisit the approach of [ BV16 ] . They use the resolution of the weighted sum problem as sampling device for a heuristic to find the whole Pareto front. Basically, they provide a QAOA circuit with pre-trained variational parameters that work for arbitrary weighted sum weights. By executing this circuit for randomly sample weights and filtering the efficient solutions, they show that this approach is compatible to some classical multicriteria solvers in speed and accuracy. 1.2. Our Contribution By investigating the algorithm presented in [ EWS24 ] , we introduce