Pith. sign in

REVIEW 4 major objections 5 minor 59 references

A Quantum Genetic Algorithm Framework for the MaxCut Problem

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper proposes a quantum genetic algorithm that encodes all MaxCut partitions in superposition, amplifies high-fitness cuts with Grover search, and uses divide-and-conquer contraction, reporting exact optimal values on complete…

desk verdict A legitimate new combination of RQGA, Grover, and graph contraction for MaxCut, but the oracle threshold as written marks no states, so the reported results cannot follow from the described circuit. read the letter →

arxiv 2501.01058 v1 pith:EQMKW4AT submitted 2025-01-02 quant-ph cs.ETcs.PF

Paulo A. Viana , Fernando M. de Paula Neto This is my paper · ORCID
classification quant-phcs.ETcs.PF
keywords MaxCutquantumgeneticalgorithmGrover'sdivideandconquergraphcontractionsemidefiniteprogrammingcombinatorialoptimizationNISQ
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to show that MaxCut, a hard combinatorial optimization problem, can be handled by a quantum genetic algorithm that replaces classical crossover and mutation with Grover search. It encodes all candidate partitions in a superposition, computes each cut's value in a quantum register, amplifies the best cuts, and uses graph partitioning and contraction to fit large graphs onto a small quantum register. On complete graphs the reported method returns the true optimal cut for every tested instance, and on Erdős–Rényi random graphs it produces median solutions within 92–96% of the semidefinite programming (SDP) baseline, with best runs often reaching or exceeding SDP. A sympathetic reader would take the paper's contribution to be a concrete, near-term-hardware template for turning Grover search into a reusable optimizer rather than a proof of quantum advantage over classical heuristics.

What carries the argument

The load-bearing object is the Grover search circuit, specifically the oracle-diffuser pair. The authors define an oracle that marks basis states by the condition $f(u) > T$, where $f(u)$ is the cut value stored in the fitness register and $T$ is said to be the total number of edges $|E|$; since $f(u) \le |E|$ for every cut, this threshold as written marks no states, so the convergence mechanism in the paper needs a different threshold to be well-defined. The fitness values are computed by a reversible circuit built from CNOT and Toffoli gates, and the comparison $f(u)>T$ is implemented with a quantum ripple-carry adder. Around this core, the algorithm wraps a divide-and-conquer layer: graph partitioning into subgraphs, independent local QGA optimization, and graph contraction into a meta-graph whose edge weights encode the boundary connections, using the $\mathbb{Z}_2$ symmetry of MaxCut to make subgraph flips harmless.

What would settle it

Build the Section 4.6 oracle for any small graph, set the threshold to $T=|E|$, and enumerate the truth table of $f(u)>T$ over all $2^{|V|}$ partitions: the predicate is false for every partition, so a complete Grover iteration leaves the uniform superposition unchanged and measuring gives no better than a random cut, directly contradicting the reported convergence.

Watch

Extended reading notes

Core claim

At the paper's center is the claim that MaxCut can be attacked by a quantum genetic algorithm in which no genetic operators are needed. All candidate partitions are loaded into a single superposition, a reversible fitness circuit computes each partition's cut value into a fitness register, and a Grover oracle built from a quantum ripple-carry adder flips the phase of states whose cut value exceeds a threshold $T$. Repeated Grover diffusion then amplifies those high-fitness states, and a measurement returns a good partition. To handle graphs larger than the available qubits, the algorithm recursively partitions the graph, solves each subgraph independently, and contracts the subgraphs into a weighted meta-graph, exploiting the $\mathbb{Z}_2$ symmetry of MaxCut so a subgraph solution and its complement are treated as equivalent. The paper reports that on complete graphs this procedure returns the exact MaxCut $\lfloor n^2/4 \rfloor$ for every tested size, while SDP gives at most about 99.7% there, and on Erdős–Rényi graphs the median QGA value is 92–96% of the SDP value, with the best of several runs often matching or beating SDP.

Load-bearing premise

The load-bearing premise is that the oracle marks a non-empty set of high-fitness states, yet the stated threshold $T=|E|$ makes the marked set empty because every cut value is at most $|E|$.

Editorial extensions

If this is right

  • If the reported complete-graph results hold, the QGA would outperform the SDP baseline on the tested range, returning exact cut values while SDP saturates near 99.7%.
  • The $O(\sqrt{2^M})$ query count gives a quadratic speedup over exhaustive enumeration of cuts, although the search space itself is exponential in the number of vertices.
  • The divide-and-conquer contraction makes the algorithm executable with a small qubit register on larger graphs, at the cost of losing boundary edges between subgraphs.
  • On Erdős–Rényi graphs, median values within 92–96% of SDP and best runs at or above SDP suggest the framework is competitive as a heuristic, not just as an exact solver for complete graphs.
  • The same oracle-plus-contraction skeleton could be adapted to other binary assignment problems, as the conclusion suggests for weighted MaxCut and related combinatorial problems.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A corrected threshold would make the Grover mechanism concrete: shifting $T$ from $|E|$ to something like $\lfloor |E|/2 \rfloor$ (or a dynamic value) gives a non-empty marked set and makes the claimed convergence testable on small simulators.
  • The exactness on complete graphs is a special case of the contraction heuristic: because complete graphs have uniform boundary structure, merging subgraphs costs no edges, whereas on sparse graphs boundary-edge loss is real, so the 92–96% median on Erdős–Rényi graphs is the more typical operating regime.
  • Section 4.8's resource formula adds the Grover iteration count $m$ to the qubit count as if it were a register dimension; correcting that conflation would change the reported scaling and should be settled before using the complexity result for hardware estimates.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This paper proposes a quantum genetic algorithm (QGA) for MaxCut. The algorithm encodes vertex partitions in a superposition, computes cut values into a fitness register via a unitary operator, applies an oracle that flips phases of states with fitness above a threshold T, and uses Grover diffusion to amplify those states. To scale beyond the available qubit register, the authors propose a divide-and-conquer heuristic that partitions graphs, solves subgraphs with the QGA, and merges solutions by graph contraction. The paper claims exact optimal MaxCut values on complete graphs, median solutions at 92--96% of the SDP values on Erdős--Rényi random graphs, and a Grover query complexity of O(sqrt(2^{M-1})). A GitHub repository and SDP comparisons are provided.

Significance. If the mechanism were correct, the paper would offer an interesting NISQ-era heuristic that combines Grover-style amplitude amplification with divide-and-conquer for MaxCut, with potential generalization to other combinatorial problems. The use of the analytically known complete-graph optimum and the direct comparison against SDP are appropriate validation choices, and the paper is honest about the loss of boundary edges in graph contraction. However, the central mechanism is internally inconsistent as written: the oracle threshold of |E| marks no states, so the reported results cannot follow from the described circuit. The complexity analysis also conflates search-space dimension, fitness-register width, and qubit count. These are load-bearing defects, not presentation issues, and they invalidate the empirical and theoretical claims in the current form.

major comments (4)
  1. [Section 4.6 and Algorithm 1] The oracle is defined with T = |E| and g(f(u),T) = 1 iff f(u) > T. Since every cut value satisfies f(u) <= |E|, and f(u) = |E| is possible only when every edge is cut (i.e., the graph is bipartite and the cut is a proper 2-coloring), the marked set is empty for the complete graphs and Erdős--Rényi graphs in Tables 1--3. For example, K3 has max cut 2 < 3 = |E|. With zero marked states, Grover iteration leaves the uniform superposition unchanged up to a global phase, so measurements sample all partitions uniformly and are independent of their cut values. Consequently the abstract's claims and the results in Tables 1--3 cannot follow from the described algorithm. The paper must either use a threshold that is actually attainable and below the maximum cut value, consistent with the lower-bound discussion in Section 4.4, or specify an adaptive threshold scheme such as Durr--Hoyer maximum finding; as written, Section 4.6 contradicts the algorithm's stated convergence behavior.
  2. [Algorithm 1, step 6, and Section 4.8] The query complexity is derived as O(sqrt(2^{M-1})) by writing N = 2^M / 2. This conflates the fitness-register width M with the search space. The search is over vertex partitions, whose number is 2^{|V|}; the factor 2 for 'two exact solutions' would affect the number of marked states, not the dimension of the fitness register. With exactly two marked states, the Grover iteration count should be O(sqrt(2^{|V|}/2)), not O(sqrt(2^{M-1})). As written, the claimed complexity is independent of the number of vertices and is therefore not supported.
  3. [Section 4.8] The resource formula f(|V|,n,M,m) = |V|*n + 2*(M+m) + 3 treats the Grover iteration count m as an additive number of qubits. Grover iterations are repeated applications of the same oracle and diffusion circuits; they increase circuit depth, not qubit width. The formula also appears to insert m = O(sqrt(2^M)) as though it were a register size. The derived bound g(n) is garbled: a complete graph has |E| = n(n-1)/2, so M = ceil(log2(n(n-1)/2)), not the expression used in the manuscript. This invalidates the stated minimum-qubit-requirement claim.
  4. [Sections 4.2.1 and 5.4] The convergence loop is not specified. The text says the cycle is repeated until the threshold ceases to improve, but Algorithm 1 has no threshold update rule, stopping criterion, or specification of how many repetitions and measurement shots are used. Table 3 reports 'best picks after multiple runs' without defining the number of runs or the selection procedure. Under the circuit as described, whose actual behavior is uniform sampling, a 'best of many runs' is just random search. Without these details, the empirical tables cannot be interpreted as evidence for the proposed algorithm.
minor comments (5)
  1. [References] There are duplicated references: the Goemans--Williamson SDP paper appears as [10], [11], and [12], Shor appears as both [45] and [46], and Nielsen--Chuang appears as both [33] and [34]. The bibliography should be consolidated.
  2. [Throughout] There are numerous typographical and rendering errors, including 'tThen' in Section 4.2.1, 'Erd˝os-R´enyi' in the abstract, 'respectivelly' and 'proposted' in Section 5, and 'withing the theoretical limit' in Section 5.3.
  3. [Section 4.5] The function signature f: {G,K} -> N is unclear; the fitness function should be defined on the space of partitions, e.g., {0,1}^{|V|}, rather than on the unexplained symbols G and K.
  4. [Section 2.1] The CNOT matrix and behavior are referenced with 'Figure ??', indicating a missing or unresolved cross-reference.
  5. [Section 5.1] The experimental setup does not report the number of runs, number of measurement shots, random seeds, simulator version, or details of the METIS partitioning and boundary-edge handling, which are needed for reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the QGA's empirical claims rest on an external benchmark (SDP and the analytic complete-graph optimum) and on a standard Grover-based search, not on fitted inputs or a self-citation chain.

full rationale

No circular derivation is present in the paper's central chain. The fitness operator U_f computes MaxCut values directly from the graph and a candidate partition, and the oracle/diffusion steps are the standard Grover search primitive; no parameter is fitted to a subset of the reported data and then renamed as a prediction. The complete-graph optimum floor(n^2/4) is used only as an independent benchmark in Table 1, not as an input that defines the QGA output. The cited RQGA framework [51] and the divide-and-conquer idea [56] are prior external work by other groups, and they are used as framework references rather than as load-bearing self-citations that force the present result. The most serious issue noted by the reader — that the Section 4.6 oracle threshold T=|E| makes the marked set empty because f(u) is bounded above by |E| — is an internal correctness/implementation flaw in the circuit description, not a circularity: the reported results are not equivalent to the oracle's inputs by construction. The derivation is therefore self-contained with respect to the circularity criteria.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim rests on several hand-chosen thresholds and implicit assumptions about oracle behavior and reversible fitness circuits. No new physical entities are introduced. The most consequential parameter is the oracle threshold T = |E|, which as written makes the marked set empty.

free parameters (4)
  • Oracle threshold T = |E| (number of edges)
    Set by hand in Section 4.6. With the marking condition f(u) > T, no state qualifies, so the algorithm's search behavior depends critically on this arbitrary and unreachable threshold.
  • Subgraph size limit n = not specified in the paper
    Section 4.7 partitions until |Vi| <= n, where n is the available qubit count. This controls how many boundary edges are lost and therefore affects solution quality, but the value used in experiments is not reported.
  • Lower-bound cutoff = 1/2 of total edge weight
    Sections 4.4 and 4.5 exclude individuals with cut <= 1/2 |E|. This is an ad hoc design choice that reshapes the search space and is not derived from optimality conditions.
  • Number of Grover iterations = O(sqrt(2^M))
    Mentioned in Section 4.8 as m = O(sqrt(2^M)). The exact integer iteration count used in the experiments is not given, and it is unclear whether the threshold or marked set changes between iterations.
assumptions (4)
  • domain assumption Grover's algorithm can be applied with an oracle that marks all states above a fixed threshold, with a single search procedure.
    Invoked in Sections 4.2.3 and Algorithm 1. Standard Grover search assumes a known number of marked states and an oracle that correctly identifies them; here the marked set is unknown and the threshold is fixed, so the quadratic speedup claim does not follow without modification.
  • domain assumption The MaxCut fitness function can be implemented as a reversible quantum circuit with manageable overhead.
    Section 4.5 says Ufit is built from CNOT and Toffoli gates but gives no gate count or circuit size. The complexity analysis in Section 4.8 ignores the cost of evaluating the fitness function, which could be O(|E|) gates per query.
  • domain assumption Graph contraction after independent subgraph optimization preserves enough boundary information to reconstruct a high-quality global cut.
    Section 4.7 explicitly lists loss of boundary edges as a limitation, yet the empirical claims rely on this heuristic. The paper assumes this loss is acceptable for complete and random graphs.
  • standard math Standard quantum information background, including unitarity, no-cloning, and tensor product formalism.
    Used throughout Section 2 without proof; these are standard background results from quantum computing.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Quantum Genetic Algorithm Framework for the MaxCut Problem." pith.science (2026). https://pith.science/paper/EQMKW4AT

@misc{pith2026250101058,
  author       = {Pith},
  title        = {Pith review of: A Quantum Genetic Algorithm Framework for the MaxCut Problem},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EQMKW4AT}},
  note         = {Machine review of arXiv:2501.01058}
}
read the original abstract

The MaxCut problem is a fundamental problem in Combinatorial Optimization, with significant implications across diverse domains such as logistics, network design, and statistical physics. The algorithm represents innovative approaches that balance theoretical rigor with practical scalability. The proposed method introduces a Quantum Genetic Algorithm (QGA) using a Grover-based evolutionary framework and divide-and-conquer principles. By partitioning graphs into manageable subgraphs, optimizing each independently, and applying graph contraction to merge the solutions, the method exploits the inherent binary symmetry of MaxCut to ensure computational efficiency and robust approximation performance. Theoretical analysis establishes a foundation for the efficiency of the algorithm, while empirical evaluations provide quantitative evidence of its effectiveness. On complete graphs, the proposed method consistently achieves the true optimal MaxCut values, outperforming the Semidefinite Programming (SDP) approach, which provides up to 99.7\% of the optimal solution for larger graphs. On Erd\H{o}s-R\'{e}nyi random graphs, the QGA demonstrates competitive performance, achieving median solutions within 92-96\% of the SDP results. These results showcase the potential of the QGA framework to deliver competitive solutions, even under heuristic constraints, while demonstrating its promise for scalability as quantum hardware evolves.

Figures

Figures reproduced from arXiv: 2501.01058 by the authors.

Figure 1
Figure 1. Diagram of the proposed framework. 4.6 Oracle Subcircuit The oracle subcircuit marks high-fitness solutions by using a boolean function and phase￾kickback to evaluates and flip the phase of states. Afterwards, those marked states will be amplified by Grover’s Algorithm. The Oracle circuit is implemented to perform the transformation: O : |u⟩ ⊗ |f(u)⟩ → (−1)g(f(u),T) |u⟩ ⊗ |f(u)⟩, where: • |u⟩: Represents the quantum… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 48 canonical work pages

  1. [1]

    Graph coloring using the reduced quantum genetic algorithm

    S. M. Ardelean and M. Udrescu. “Graph coloring using the reduced quantum genetic algorithm”. In: PeerJ Computer Science 7 (2022). doi: 10.7717/peerj-cs.836

    work page doi:10.7717/peerj-cs.836 2022

  2. [2]

    Machine Learning for Com- binatorial Optimization: A Methodological Tour d’Horizon

    Yoshua Bengio, Andrea Lodi, and Antoine Prouvost. “Machine Learning for Com- binatorial Optimization: A Methodological Tour d’Horizon”. In: arXiv preprint arXiv:1811.06128 (Mar. 2020)

    arXiv 2020

  3. [3]

    Quantum machine learning

    Jacob Biamonte et al. “Quantum machine learning”. In: Nature 549.7671 (2017), pp. 195–202

    work page 2017

  4. [4]

    Combinatorial optimization and reasoning with graph neu- ral networks

    Quentin Cappart et al. “Combinatorial optimization and reasoning with graph neu- ral networks”. In: arXiv preprint arXiv:2102.09544 (Apr. 2021). 20

    arXiv 2021

  5. [5]

    Cuccaro et al

    Steven A. Cuccaro et al. A new quantum ripple-carry addition circuit . 2004. arXiv: quant - ph / 0410184 [quant-ph]. url: https : / / arxiv . org / abs / quant - ph / 0410184

    work page 2004

  6. [6]

    Rapid solution of problems by quantum com- putation

    David Deutsch and Richard Jozsa. “Rapid solution of problems by quantum com- putation”. In: Proceedings of the Royal Society of London. Series A: Mathematical and Physical Sciences 439.1907 (1992), pp. 553–558

    work page 1992

  7. [7]

    Some extremal properties of bipartite subgraphs

    C. S. Edwards. “Some extremal properties of bipartite subgraphs”. In: Canadian Journal of Mathematics 25.3 (1973), pp. 475–485. doi: 10.4153/cjm-1973-048-x

    work page doi:10.4153/cjm-1973-048-x 1973

  8. [8]

    A. E. Eiben and J. E. Smith. Introduction to evolutionary computing. Natural Com- puting Series. 2015. doi: 10.1007/978-3-662-44874-8

    doi:10.1007/978-3-662-44874-8 2015

Show all 59 references
  1. [9]

    A review on quantum search algo- rithms

    Pulak Ranjan Giri and Vladimir E Korepin. “A review on quantum search algo- rithms”. In: Quantum Information Processing 16 (2017), pp. 1–36

    2017

  2. [10]

    Improved approximation algorithms for maximum cut and satisfiability problems using semidefinite programming

    Michel X Goemans and David P Williamson. “Improved approximation algorithms for maximum cut and satisfiability problems using semidefinite programming”. In: Journal of the ACM (JACM) 42.6 (1995), pp. 1115–1145

    1995

  3. [11]

    879-approximation algorithms for MAX CUT and MAX 2SAT

    Michel X. Goemans and David P. Williamson. “.879-approximation algorithms for MAX CUT and MAX 2SAT”. In: Proceedings of the Twenty-Sixth Annual ACM Symposium on Theory of Computing, STOC ’94 . New York, NY, USA: Association for Computing Machinery, May 1994, pp. 422–431

    1994

  4. [12]

    Improved approximation algorithms for maximum cut and satisfiability problems using semidefinite programming

    Michel X. Goemans and David P. Williamson. “Improved approximation algorithms for maximum cut and satisfiability problems using semidefinite programming”. In: J. ACM 42.6 (Nov. 1995), pp. 1115–1145

    1995

  5. [13]

    Quantum carry-save arithmetic

    P. Gossett. “Quantum carry-save arithmetic”. In: (1998). arXiv: quant-ph/9808061

    1998 arXiv

  6. [14]

    A fast quantum mechanical algorithm for database search

    L. K. Grover. “A fast quantum mechanical algorithm for database search”. In: Proceedings of the 28th Annual ACM Symposium on Theory of Computing . 1996

    1996

  7. [15]

    A fast quantum mechanical algorithm for database search

    Lov K Grover. “A fast quantum mechanical algorithm for database search”. In: Proceedings of the twenty-eighth annual ACM symposium on Theory of computing . 1996, pp. 212–219

    1996

  8. [16]

    Qaoa for Max-cut requires hundreds of qubits for quantum speed-up

    G. G. Guerreschi and A. Y. Matsuura. “Qaoa for Max-cut requires hundreds of qubits for quantum speed-up”. In: Scientific Reports 9.1 (2019). doi: 10 . 1038 / s41598-019-43176-9

    2019

  9. [17]

    A more efficient deterministic annealing neu- ral network algorithm for the max-bisection problem

    Shicong Jiang and Chuangyin Dang. “A more efficient deterministic annealing neu- ral network algorithm for the max-bisection problem”. In: Neurocomputing 458 (2021), pp. 428–439

    2021

  10. [18]

    Local optimization and the Traveling Salesman Problem

    David S. Johnson. “Local optimization and the Traveling Salesman Problem”. In: Automata, Languages and Programming. Ed. by Michael S. Paterson. Lecture Notes in Computer Science. Berlin, Heidelberg: Springer, 1990, pp. 446–461

    1990

  11. [19]

    On the role of entanglement in quantum-computational speed-up

    Richard Jozsa and Noah Linden. “On the role of entanglement in quantum-computational speed-up”. In: Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences 459.2036 (2003), pp. 2011–2032

    2003

  12. [20]

    Combinatorial Algorithms for Portfolio Optimization Problems – Case of Risk Moderate Investor

    A. Juarna. “Combinatorial Algorithms for Portfolio Optimization Problems – Case of Risk Moderate Investor”. In: J. Phys.: Conf. Ser. 820 (Mar. 2017), p. 012028. 21

    2017

  13. [21]

    Chapter 4 The travel- ing salesman problem

    Michael J¨ unger, Gerhard Reinelt, and Giovanni Rinaldi. “Chapter 4 The travel- ing salesman problem”. In: Handbooks in Operations Research and Management Science, volume 7 of Network Models . Elsevier, Jan. 1995, pp. 225–330

    1995

  14. [22]

    Reducibility among combinatorial problems

    R. M. Karp. “Reducibility among combinatorial problems”. In: Complexity of Com- puter Computations . 1972, pp. 85–103. doi: 10.1007/978-1-4684-2001-2_9

    1972 doi

  15. [23]

    P. Kaye, R. Laflamme, and M. Mosca. An Introduction to Quantum Computing

  16. [24]

    A comparison study of harmony search and genetic algorithm for the max-cut problem

    Yong-Hyuk Kim, Yourim Yoon, and Zong Woo Geem. “A comparison study of harmony search and genetic algorithm for the max-cut problem”. In: Swarm and evolutionary computation 44 (2019), pp. 130–135

    2019

  17. [25]

    Optimization by Simulated An- nealing

    S. Kirkpatrick, C. D. Gelatt, and M. P. Vecchi. “Optimization by Simulated An- nealing”. In: Science 220.4598 (May 1983), pp. 671–680

    1983

  18. [26]

    Kitaev, A

    A. Kitaev, A. Shen, and M. Vyalyi. Classical and quantum computation . Graduate Studies in Mathematics. 2002. doi: 10.1090/gsm/047

    2002 doi

  19. [27]

    Combinatorial Optimization: Theory and Algo- rithms

    Bernhard Korte and Jens Vygen. Combinatorial Optimization: Theory and Algo- rithms. 3rd. Algorithms and Combinatorics. Springer-Verlag, Berlin Heidelberg, 2006

    2006

  20. [28]

    Quantum genetic algorithms for computer scientists

    Rafael Lahoz-Beltra. “Quantum genetic algorithms for computer scientists”. In: Computers 5.4 (2016), p. 24

    2016

  21. [29]

    Quantum Algorithms for Combinatorial Optimization

    John Martin. “Quantum Algorithms for Combinatorial Optimization”. In: Quantum Computing Review 5 (2022), pp. 123–145

    2022

  22. [30]

    Reinforcement learning for combinatorial optimization: A survey

    Nina Mazyavkina et al. “Reinforcement learning for combinatorial optimization: A survey”. In: Computers & Operations Research 134 (Oct. 2021), p. 105400

    2021

  23. [31]

    An experimental evaluation of semidefinite pro- gramming and spectral algorithms for Max Cut

    R. Mirka and D. P. Williamson. “An experimental evaluation of semidefinite pro- gramming and spectral algorithms for Max Cut”. In: ACM Journal of Experimental Algorithmics 28 (2023), pp. 1–18. doi: 10.1145/3609426

    2023 doi

  24. [32]

    Multivalued logic gates for quantum compu- tation

    A. Muthukrishnan and C. R. Stroud. “Multivalued logic gates for quantum compu- tation”. In: Physical Review A 62.5 (2000). doi: 10.1103/physreva.62.052309

    2000 doi

  25. [33]

    M. A. Nielsen and I. L. Chuang. Quantum Computation and Quantum Information

  26. [34]

    Quantum Computation and Quantum In- formation

    Michel A Nielsen and Isaac L Chuang. Quantum Computation and Quantum In- formation. Cambridge University Press, 2000

    2000

  27. [35]

    A practical architecture for reliable quantum computers

    M. Oskin, F. Chong, and I. Chuang. “A practical architecture for reliable quantum computers”. In: IEEE Computer 35.1 (2002), pp. 79–87

    2002

  28. [36]

    B. Parhami. Computer Arithmetic: Algorithms and Hardware Designs . Oxford: Ox- ford University Press, 2000

    2000

  29. [37]

    Clustering pairwise distances with missing data: Maximum cuts versus normalized cuts

    Jan Poland and Thomas Zeugmann. “Clustering pairwise distances with missing data: Maximum cuts versus normalized cuts”. In: Lecture Notes in Computer Sci- ence (2006), pp. 197–208. doi: 10.1007/11893318_21

    2006 doi

  30. [38]

    Maximum cuts and large bipartite subgraphs

    Svatopluk Poljak and Tuza. “Maximum cuts and large bipartite subgraphs”. In: DIMACS Series in Discrete Mathematics and Theoretical Computer Science (June 1995), pp. 181–244. 22

    1995

  31. [39]

    Self-repairing embryonic memory ar- rays

    L. Prodan, M. Udrescu, and M. Vladutiu. “Self-repairing embryonic memory ar- rays”. In: Proc. IEEE NASA/DoD Conference on Evolvable Hardware . Seattle, 2004, pp. 130–137

    2004

  32. [40]

    Survivability of embryonic memories: Analysis and design principles

    L. Prodan, M. Udrescu, and M. Vladutiu. “Survivability of embryonic memories: Analysis and design principles”. In: Proc. IEEE NASA/DoD Conference on Evolv- able Hardware (EH’05) . 2005, pp. 280–289

    2005

  33. [41]

    Traveling salesman problem heuristics: Leading methods, im- plementations and latest advances

    C´ esar Rego et al. “Traveling salesman problem heuristics: Leading methods, im- plementations and latest advances”. In: European Journal of Operational Research 211.3 (June 2011), pp. 427–441

    2011

  34. [42]

    Computational complexity, genetic program- ming, and implications

    B. Rylander, T. Soule, and J. Foster. “Computational complexity, genetic program- ming, and implications”. In: Proc. 4th EuroGP. 2001, pp. 348–360

    2001

  35. [43]

    Quantum evolutionary programming

    B. Rylander et al. “Quantum evolutionary programming”. In: Proceedings of the Ge- netic and Evolutionary Computation Conference (GECCO-2001) . 2001, pp. 1005– 1011

    2001

  36. [44]

    Combinatorial optimization and Green Logistics

    Abdelkader Sbihi and Richard W. Eglese. “Combinatorial optimization and Green Logistics”. In: 4OR 5.2 (July 2007), pp. 99–116

    2007

  37. [45]

    Algorithms for quantum computation: Discrete logarithms and factor- ing

    P. W. Shor. “Algorithms for quantum computation: Discrete logarithms and factor- ing”. In: Proc. 35th Symposium on Foundations of Computer Science. 1994, pp. 124– 134

    1994

  38. [46]

    Algorithms for quantum computation: discrete logarithms and fac- toring

    Peter W Shor. “Algorithms for quantum computation: discrete logarithms and fac- toring”. In: Proceedings 35th annual symposium on foundations of computer science. Ieee. 1994, pp. 124–134

    1994

  39. [47]

    L. Spector. Automatic Quantum Computer Programming: A Genetic Programming Approach. Boston: Kluwer Academic Publishers, 2004

    2004

  40. [48]

    Genetic programming for quantum computers

    L. Spector, H. Barnum, and H. Bernstein. “Genetic programming for quantum computers”. In: Genetic Programming 1998: Proceedings of the Third Annual Con- ference. Madison, Wisconsin, 1998, pp. 365–373

    1998

  41. [49]

    Finding a better-than-classical quantum and/or algorithm using genetic programming

    L. Spector et al. “Finding a better-than-classical quantum and/or algorithm using genetic programming”. In: Proceedings of 1999 Congress of Evolutionary Compu- tation. Piscataway, NJ: IEEE, 1999, pp. 2239–2246

    1999

  42. [50]

    Quantum computing applications of genetic programming

    L. Spector et al. “Quantum computing applications of genetic programming”. In: Advances in Genetic Programming 3.7 (1998), pp. 135–160

    1998

  43. [51]

    Implementing Quantum Genetic al- gorithms

    M. Udrescu, L. Prodan, and M. Vladutiu. “Implementing Quantum Genetic al- gorithms”. In: Proceedings of the 3rd Conference on Computing Frontiers . 2006, pp. 71–82. doi: 10.1145/1128022.1128034

    2006

  44. [52]

    Quantum Networks for elementary arith- metic operations

    V. Vedral, A. Barenco, and A. Ekert. “Quantum Networks for elementary arith- metic operations”. In: Physical Review A 54.1 (1996), pp. 147–153. doi: 10.1103/ physreva.54.147

    1996

  45. [53]

    Improved quantum ripple-carry addition circuit

    F. Wang et al. “Improved quantum ripple-carry addition circuit”. In: Science China Information Sciences 59.4 (2016). doi: 10.1007/s11432-015-5411-x

    2016 doi

  46. [54]

    An improved discrete Hopfield neural network for Max-Cut prob- lems

    Jiahai Wang. “An improved discrete Hopfield neural network for Max-Cut prob- lems”. In: Neurocomputing 69.13-15 (2006), pp. 1665–1669. 23

    2006

  47. [55]

    Quantum-inspired evolutionary algorithms: A survey and empirical study

    G. Zhang. “Quantum-inspired evolutionary algorithms: A survey and empirical study”. In: Journal of Heuristics 17.3 (2010), pp. 303–351. doi: 10.1007/s10732- 010-9136-0

    2010 doi

  48. [56]

    Qaoa-in-QAOA: Solving large-scale MaxCut problems on small quantum machines

    Z. Zhou et al. “Qaoa-in-QAOA: Solving large-scale MaxCut problems on small quantum machines”. In: Physical Review Applied 19.2 (2023). doi: 10 . 1103 / physrevapplied.19.024027

    2023

  49. [57]

    Adaptive quantum approximate optimization algorithm for solving combinatorial problems on a quantum computer

    Linghua Zhu et al. “Adaptive quantum approximate optimization algorithm for solving combinatorial problems on a quantum computer”. In: Physical Review Re- search 4.3 (2022), p. 033029. 24

    2022

  50. [2006]

    doi: 10.1093/oso/9780198570004.001.0001

  51. [2012]

    doi: 10.1017/cbo9780511976667

    doi

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.