Pith. sign in

REVIEW 4 major objections 4 minor 64 references

Reinforcement learning Based Automated Design of Differential Evolution Algorithm for Black-box Optimization

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

Pith's one-line read An RL agent can auto-design differential evolution for unseen black-box problems.

desk verdict Offline, per-problem DE design via RL and ELA is genuinely new, but zero-variance results and an undefined AEI make the empirical headline unsupportable. read the letter →

arxiv 2501.12881 v1 pith:F4RBZJ6U submitted 2025-01-22 cs.NE cs.AI

classification cs.NEcs.AI
keywords differentialevolutionreinforcementlearningblack-boxoptimizationexploratorylandscapeanalysismeta-learningdeepQ-networkalgorithmdesignBBOB2009
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

This paper proposes a framework, rlDE, in which a reinforcement-learning agent learns, offline, to assemble a complete differential evolution algorithm for a given black-box optimization problem. The agent reads a problem through 62 exploratory landscape analysis features—cheap statistics computed once from a single random sample—and outputs a configuration that includes an initialization strategy, a mutation strategy, a crossover strategy, and values for population size, scale factor, and crossover rate. The paper's claim is that this learned mapping, trained on a majority of the BBOB2009 noiseless functions, generalizes to unseen problems: on six held-out functions the resulting algorithm attains the best aggregated evaluation indicator (AEI 18.00) among all compared RL-assisted DE variants and is competitive with the best traditional DEs. If the claim holds, it would mean that algorithm design for black-box optimization can be automated by learning the relationship between cheap landscape statistics and effective algorithm configurations, rather than relying on expert knowledge or per-problem search.

What carries the argument

The load-bearing mechanism is the multi-armed-bandit deep Q-network (MADQN), which extends DDQN so that a single state produces several Q-value vectors, one per design object (initialization, mutation, crossover, and control parameters), letting the agent choose multiple strategy types simultaneously. The state is provided by exploratory landscape analysis: 62 features computed from one random initial sample, covering y-distribution, levelset classification errors, linear-model fits, funnel-structure statistics, information content, dispersion, and PCA of the decision space, plus the problem dimension. The design space being searched is a subset of 40 strategy combinations, with population size among $\{5D,7D,9D,11D,13D\}$, scale factor in steps of $0.05$, and crossover rate in steps of $0.1$.

What would settle it

A decisive check is to grid-search the same design space on each of the 24 noiseless BBOB2009 functions and compare the best per-function configuration with the one the agent chooses from the 62-feature state; if functions that are close in feature space have very different best configurations, or if the agent's choices are consistently worse than the oracle, the state representation is insufficient to support the claimed mapping.

Watch

Extended reading notes

Core claim

The central discovery the paper argues for is that the relationship between a black-box problem's landscape and a well-performing differential evolution configuration can be meta-learned by a double deep Q-network. The agent is trained on triples of landscape-feature state, generated algorithm action, and observed performance reward, where the reward is simply $r_t = e^{-f_t^*}$ with $f_t^*$ the best objective value found. After training, the agent is frozen and applied to an unseen problem by computing its 62 features once and decoding the network's output into a full DE variant. On the six BBOB2009 test functions selected, the paper reports zero standard deviation across 31 runs for each function and an aggregated evaluation indicator of 18.00, ranking first among the compared algorithms; this is offered as evidence that the generated algorithms are stable and effective.

Load-bearing premise

The load-bearing premise is that 62 static features, computed once from a single random sample, capture enough of a problem's structure to determine which differential evolution configuration will perform best.

Editorial extensions

If this is right

  • For a previously unseen problem, rlDE produces a complete DE without any per-problem search, so the cost of algorithm design becomes one round of feature computation plus one network forward pass.
  • The learned agent outperforms RL-assisted DE variants that adapt operators during evolution, which suggests that offline design from problem features can be more effective than online adaptation from population statistics.
  • If the mapping generalizes beyond BBOB2009, the same meta-learning scheme could be applied to other evolutionary algorithms, since the framework only needs a parametrized algorithm family and a performance signal.
  • The reported zero variance across 31 runs implies the agent converges to a single deterministic configuration per test function, making the generated algorithms repeatable.

Reading between the lines

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

  • A direct test of the representation would be to compare the agent's chosen configuration against a per-problem oracle that grid-searches the same 40-combination design space; the gap between them measures how much information the 62 features actually carry.
  • The framework could be extended to a continuous design space, or to a reward that tracks entire convergence curves rather than the final objective, which might improve generalization when evaluation budgets vary.
  • Because the state is computed once at the start, the approach is insensitive to the search dynamics; a natural next step is to make part of the state update during evolution, though this would trade away the cost advantage the paper argues for.
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 / 4 minor

Summary. The manuscript introduces rlDE, a reinforcement-learning framework that uses exploratory landscape analysis (ELA) features as states and a double deep Q-network (DDQN) to automatically select initialization, mutation, crossover, and control parameters for differential evolution (DE). The agent is trained offline on a subset of BBOB2009 problems and then generates a tailored DE configuration for unseen problems. The main empirical claim is that rlDE outperforms four RL-assisted DE variants and four traditional DE variants on six BBOB2009 test functions, achieving the best aggregated evaluation indicator (AEI 18.00) and exhibiting zero standard deviation across 31 runs. The paper also discusses the motivation for using problem characteristics instead of population characteristics and compares the framework with hyper-heuristics and self-adaptive methods.

Significance. The idea of replacing per-generation adaptive operator selection with one-shot, problem-characteristic-based DE design is timely and aligns with the growing interest in automated algorithm design and meta-learning for black-box optimization. The paper's conceptual separation of the learning episode from the using episode, and its use of ELA features as state representations, are reasonable design choices that could inspire follow-up work. However, the headline empirical result—that rlDE ranks first with AEI 18.00—is not supported by the reported data: the zero-variance results in Table III are implausible for a stochastic algorithm, the AEI formula in Eq. (10) becomes undefined when the standard deviation is zero, and the claimed design-space size in the abstract is inconsistent with Table II. These issues are load-bearing because the paper's central claim rests on the quantitative comparison. If the experiments are rerun with proper independent seeded runs and the statistical analysis is repaired, the framework may still be a valuable proof-of-concept; as presented, the evidence for superiority is not credible.

major comments (4)
  1. [V-B, Table III] Table III reports identical worst, best, median, and mean values with standard deviation 0.0 for all six BBOB2009 test functions across 31 runs. Since rlDE uses random initialization (Table II), stochastic mutation and crossover operators, and epsilon-greedy action selection, independent runs must exhibit nonzero variance unless the runs share a random seed. If they share a seed, the 31 runs are not independent and the Wilcoxon rank-sum tests in Table IV are invalid. If they do not share a seed, the reported zero variance indicates a reporting or implementation error. The manuscript explicitly notes this lack of variability and calls it 'highly stable,' but it does not address the statistical breakdown. This directly invalidates the quantitative comparison that supports the paper's strongest claim.
  2. [V-B, Eq. (10)] The AEI calculation uses Z-score normalization Z_k* = (1/N) sum_n v_{k,n}^* / sigma_*, dividing by the per-function standard deviation over repeated runs. For rlDE, Table III reports sigma_* = 0 on every testing function, so the Z scores and hence the AEI value of 18.00 are undefined under the formula as written. The paper gives no alternative handling for zero variance, and the displayed AEI ranking therefore lacks a rigorous basis.
  3. [Abstract and IV-C, Table II] The abstract states that the proof-of-concept considers a subset of 40 possible strategy combinations and parameter optimizations, but Table II defines 5 initialization strategies x 10 mutation strategies x 2 crossover strategies x 5 population sizes x 41 scale factors x 21 crossover rates = 430,500 combinations. Moreover, Section IV-C describes an output layer with only 5 neurons, which cannot encode that design space directly. The discrepancy between the stated action space and the architecture description makes it unclear what the agent actually outputs and how the 40-combination subset (or the full space) is represented. This needs clarification and a consistent action-encoding description.
  4. [IV-C, Algorithm 5 and Section III] The state s_t is computed only from ELA features on the initial random sample of points, and the paper does not provide a concrete test of whether these static features are sufficient to determine the best DE configuration. The authors themselves note in the conclusion that the RL agent's performance depends heavily on the breadth and representativeness of training problems, but they do not present an ablation or a baseline comparison (e.g., an agent using a fixed random state, or a version without ELA). Since the state representation is the core of the claimed mapping from problem characteristics to algorithm design, an experimental comparison is needed to substantiate that the ELA features, rather than the training distribution or reward scale, drive the reported performance.
minor comments (4)
  1. [Section III] There are multiple typos in this section: 'Curvatyre' should be 'Curvature,' 'Generakized' should be 'Generalized,' 'Disperison' should be 'Dispersion,' and 'meta-modal' should be 'meta-model.' Also, the unresolved reference 'Section ??' in Section IV-C should be fixed.
  2. [Section V-B, Table IV] The table caption and the accompanying text use 'rom' instead of 'from' and 'llustrates' instead of 'illustrates' in Figure 4. The notation 'vavg(vstd)' is inconsistent with the column headers and the main text; please unify the notation for the average and standard deviation.
  3. [Section IV-B, Figure 1] Figure 1 is described as showing the overall framework, but the text does not explain the roles of the three color-coded starting points or how information flows between the RL process and the DE process. Adding a short walkthrough of the figure would improve readability.
  4. [Section II-C, Table I] The table lists DEDQN and DEDDQN as related work, but the main text does not clearly differentiate the proposed MADQN from these prior DQN-based methods beyond the multi-object action architecture. A sentence clarifying the novelty in action representation would help position the contribution.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the RL-generated DE is evaluated on held-out BBOB2009 functions, and the reward is the observed optimization performance, not a fitted constant.

full rationale

The claimed derivation chain is that ELA features computed once from an initial sample form the RL state, the DDQN agent selects a DE configuration from a predefined design space, and that configuration is scored by running the resulting DE on the sampled problem and converting the achieved objective value into reward. This is a standard offline meta-learning loop: the agent must generalize from the 18 training functions to the six held-out testing functions, and the test performance is not encoded in the state, action, or reward by construction. The reward r_t = e^{-f*_t} is the actual optimization objective being pursued, not a parameter fitted to the test results. The only author self-citations, such as Tan and Li [1] and Tan et al. [61], are used as related work and as comparison algorithms, not as load-bearing justifications for the framework's premises, so they do not make the derivation circular. The reported zero standard deviations in Table III and the consequent division by zero in Eq. (10) for the AEI indicator are serious data-integrity and statistical-validity concerns that undermine the empirical headline, but they are correctness risks rather than definitional circularity. The paper's acknowledged limitations about the predefined design space and dependence on training-problem breadth also do not reduce the central derivation to its inputs.

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

The central claim rests mainly on domain assumptions: ELA feature sufficiency, reward shaping, benchmark transferability, and the hand-picked design space. No new physical or algorithmic entities are postulated; the named MADQN is a multi-head DQN variant rather than an invented entity.

free parameters (4)
  • KDE bandwidth scaling lambda = not specified
    Controls the interval used for the number-of-peaks ELA feature in Section III; hand-set, but its value is omitted.
  • npeaks threshold = 0.1
    Threshold in the number-of-peaks calculation (Section III); hand-chosen, no sensitivity analysis.
  • DDQN training hyperparameters = not reported
    Learning rate, batch size, replay memory size, discount factor, epsilon schedule, and target update period are needed to reproduce Algorithm 5 but are not given.
  • action space subset size = 40 claimed, not enumerated
    Abstract says 40 strategy combinations, but Table II defines 5*10*2*5*41*21 combinations; the actual subset used is never specified.
assumptions (4)
  • domain assumption The 62 ELA features from a single initial sample are sufficient to predict the best DE configuration for an unseen problem.
    The state in Section IV-C contains only these features; the paper offers no evidence that static landscape features capture the algorithmic interaction that determines DE performance.
  • ad hoc to paper Reward rt = exp(-f*_t) is an adequate scalarization of DE performance.
    Eq. (8) is introduced without normalization or definition of whether f*_t is the known optimum or the best value found; most positive objectives produce near-identical rewards.
  • domain assumption Training on 18 BBOB2009 functions transfers to other black-box problems.
    The test uses six functions from the same benchmark family, so the claimed generality beyond BBOB is assumed, not demonstrated.
  • domain assumption The mu+lambda template with the listed components and discrete parameter grids is a meaningful design space for automated DE design.
    Section IV-C defines the space; it excludes modern adaptation mechanisms such as SHADE-style success-history adaptation, so 'automated design' is constrained by this hand-picked pool.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Reinforcement learning Based Automated Design of Differential Evolution Algorithm for Black-box Optimization." pith.science (2026). https://pith.science/paper/F4RBZJ6U

@misc{pith2026250112881,
  author       = {Pith},
  title        = {Pith review of: Reinforcement learning Based Automated Design of Differential Evolution Algorithm for Black-box Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F4RBZJ6U}},
  note         = {Machine review of arXiv:2501.12881}
}
read the original abstract

Differential evolution (DE) algorithm is recognized as one of the most effective evolutionary algorithms, demonstrating remarkable efficacy in black-box optimization due to its derivative-free nature. Numerous enhancements to the fundamental DE have been proposed, incorporating innovative mutation strategies and sophisticated parameter tuning techniques to improve performance. However, no single variant has proven universally superior across all problems. To address this challenge, we introduce a novel framework that employs reinforcement learning (RL) to automatically design DE for black-box optimization through meta-learning. RL acts as an advanced meta-optimizer, generating a customized DE configuration that includes an optimal initialization strategy, update rule, and hyperparameters tailored to a specific black-box optimization problem. This process is informed by a detailed analysis of the problem characteristics. In this proof-of-concept study, we utilize a double deep Q-network for implementation, considering a subset of 40 possible strategy combinations and parameter optimizations simultaneously. The framework's performance is evaluated against black-box optimization benchmarks and compared with state-of-the-art algorithms. The experimental results highlight the promising potential of our proposed framework.

Figures

Figures reproduced from arXiv: 2501.12881 by the authors.

Figure 1
Figure 1. The overall framework of rlDE, encompassing three parts: ELA, RL process and DE process, where the starting [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. MADQN process DE process executes each AlgorithmX decoded ac￾cording to at on the sampled BBOP, and the evaluation metric on AlgorithmX will be converted to rt. In this implementation, each BBOP is regarded as a minimization problem, and the evaluation metric is set to the solution quality, i.e. obtaining minimum objective values, denoted as f ∗ t . It is related to the design of reward rt. Actually, the design of t… view at source ↗
Figure 3
Figure 3. Convergence curve of normalized averaged best [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Per-instance curve of logarithm best objective with increasing costed evaluations [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: AEI values for BBOB2009 testing functions, where [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 61 canonical work pages

  1. [1]

    Differential evolution with mixed mutation strategy based on deep reinforcement learning,

    Z. Tan and K. Li, “Differential evolution with mixed mutation strategy based on deep reinforcement learning,” Applied Soft Computing, vol. 111, p. 107678, 2021

  2. [2]

    No free lunch theorems for optimization,

    D. H. Wolpert and W. G. Macready, “No free lunch theorems for optimization,” IEEE transactions on evolutionary computation, vol. 1, no. 1, pp. 67–82, 1997

  3. [3]

    A self-adapting algorithm for many-objective optimization,

    S. Reddy and G. S. Dulikravich, “A self-adapting algorithm for many-objective optimization,” Applied Soft Computing, vol. 129, p. 109484, 2022

  4. [4]

    An adaptive switching-based evolutionary algorithm for many-objective op- timization,

    S. Chen, X. Wang, J. Gao, W. Du, and X. Gu, “An adaptive switching-based evolutionary algorithm for many-objective op- timization,” Knowledge-Based Systems, vol. 248, pp. 108 915– 108 915, 2022. IEEE, VOL. X, NO. X, X X 11

  5. [5]

    Ensemble strategies for population-based optimization algorithms-a survey,

    G. Wu, R. Mallipeddi, and P. N. Suganthan, “Ensemble strategies for population-based optimization algorithms-a survey,” Swarm and Evolutionary Computation , vol. 44, pp. 695–711, 2019. [Online]. Available: https://www.sciencedirect. com/science/article/pii/S2210650217300445

  6. [6]

    Reinforcement learning-assisted evolutionary algorithm: A survey and research opportunities,

    Y. Song, Y. Wu, Y. Guo, R. Yan, P. N. Suganthan, Y. Zhang, W. Pedrycz, S. Das, R. Mallipeddi, O. S. Ajani, and Q. Feng, “Reinforcement learning-assisted evolutionary algorithm: A survey and research opportunities,” Swarm and Evolutionary Computation , vol. 86, p. 101517, 2024. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S221065...

  7. [7]

    Differential evolution based on reinforcement learning with fitness ranking for solving multimodal multiobjective problems,

    Z. Li, L. Shi, C. Yue, Z. Shang, and B. Qu, “Differential evolution based on reinforcement learning with fitness ranking for solving multimodal multiobjective problems,” Swarm and Evolutionary Computation, vol. 49, pp. 234–244, 2019

  8. [8]

    Reinforcement learning- based differential evolution for global optimization,

    I. Fister, D. Fister, and I. Fister Jr, “Reinforcement learning- based differential evolution for global optimization,” in Differ- ential Evolution: From Theory to Practice. Springer, 2022, pp. 43–75

Show all 64 references
  1. [9]

    Scheduling of continuous an- nealing with a multi-objective differential evolution algorithm based on deep reinforcement learning,

    T. Li, Y. Meng, and L. Tang, “Scheduling of continuous an- nealing with a multi-objective differential evolution algorithm based on deep reinforcement learning,” IEEE Transactions on Automation Science and Engineering, vol. 21, no. 2, pp. 1767– 1780, 2023

  2. [10]

    Improving the search perfor- mance of shade using linear population size reduction,

    R. Tanabe and A. S. Fukunaga, “Improving the search perfor- mance of shade using linear population size reduction,” in 2014 IEEE congress on evolutionary computation (CEC) . IEEE, 2014, pp. 1658–1665

  3. [11]

    il-shade: Improved l-shade algorithm for single objective real-parameter optimiza- tion,

    J. Brest, M. S. Mauˇ cec, and B. Boˇ skovi´ c, “il-shade: Improved l-shade algorithm for single objective real-parameter optimiza- tion,” in 2016 IEEE Congress on Evolutionary Computation (CEC). IEEE, 2016, pp. 1188–1195

  4. [12]

    Single objective real-parameter optimization: Algorithm jso,

    ——, “Single objective real-parameter optimization: Algorithm jso,” in 2017 IEEE congress on evolutionary computation (CEC). IEEE, 2017, pp. 1311–1318

  5. [13]

    Q-learning-based parameter control in differential evolution for structural optimization,

    T. N. Huynh, D. T. Do, and J. Lee, “Q-learning-based parameter control in differential evolution for structural optimization,” Applied Soft Computing, vol. 107, p. 107464, 2021

  6. [14]

    Rein- forcement learning-based hybrid differential evolution for global optimization of interplanetary trajectory design,

    L. Peng, Z. Yuan, G. Dai, M. Wang, and Z. Tang, “Rein- forcement learning-based hybrid differential evolution for global optimization of interplanetary trajectory design,” Swarm and Evolutionary Computation, vol. 81, p. 101351, 2023

  7. [15]

    Learning adaptive differen- tial evolution algorithm from optimization experiences by policy gradient,

    J. Sun, X. Liu, T. B¨ ack, and Z. Xu, “Learning adaptive differen- tial evolution algorithm from optimization experiences by policy gradient,” IEEE Transactions on Evolutionary Computation , vol. 25, no. 4, pp. 666–680, 2021

  8. [16]

    Variational reinforcement learning for hyper-parameter tuning of adaptive evolutionary algorithm,

    H. Zhang, J. Sun, Y. Wang, J. Shi, and Z. Xu, “Variational reinforcement learning for hyper-parameter tuning of adaptive evolutionary algorithm,” IEEE Transactions on Emerging Top- ics in Computational Intelligence , vol. 7, no. 5, pp. 1511–1526, 2022

  9. [17]

    Learning to learn evolutionary algorithm: A learnable differential evolution,

    X. Liu, J. Sun, Q. Zhang, Z. Wang, and Z. Xu, “Learning to learn evolutionary algorithm: A learnable differential evolution,” IEEE Transactions on Emerging Topics in Computational In- telligence, vol. 7, no. 6, pp. 1605–1620, 2023

  10. [18]

    Reinforcement learning-based multiobjective evolutionary algorithm for mixed- model multimanned assembly line balancing under uncertain demand,

    Z. Zhang, Q. Tang, M. Chica, and Z. Li, “Reinforcement learning-based multiobjective evolutionary algorithm for mixed- model multimanned assembly line balancing under uncertain demand,” IEEE Transactions on Cybernetics, 2023

  11. [19]

    Meta- learning in neural networks: A survey,

    T. Hospedales, A. Antoniou, P. Micaelli, and A. Storkey, “Meta- learning in neural networks: A survey,” IEEE transactions on pattern analysis and machine intelligence , vol. 44, no. 9, pp. 5149–5169, 2021

  12. [20]

    Human-level control through deep rein- forcement learning,

    V. Mnih, K. Kavukcuoglu, D. Silver, A. A. Rusu, J. Veness, M. G. Bellemare, A. Graves, M. Riedmiller, A. K. Fidjeland, G. Ostrovski et al. , “Human-level control through deep rein- forcement learning,” nature, vol. 518, no. 7540, pp. 529–533, 2015

  13. [21]

    Hyper-heuristics: A survey of the state of the art,

    E. K. Burke, M. Gendreau, M. Hyde, G. Kendall, G. Ochoa, E. ¨Ozcan, and R. Qu, “Hyper-heuristics: A survey of the state of the art,” Journal of the Operational Research Society, vol. 64, no. 12, pp. 1695–1724, 2013

  14. [22]

    Evolutionary self-adaptation: a survey of operators and strategy parameters,

    O. Kramer, “Evolutionary self-adaptation: a survey of operators and strategy parameters,” Evolutionary Intelligence, vol. 3, pp. 51–65, 2010

  15. [23]

    Self-adaptive systems: A survey of current approaches, research challenges and applications,

    F. D. Mac´ ıas-Escriv´ a, R. Haber, R. del Toro, and V. Hernandez, “Self-adaptive systems: A survey of current approaches, research challenges and applications,” Expert Systems with Applications, vol. 40, no. 18, pp. 7267–7279, 2013. [Online]. Available: https:// www.sciencedi...

  16. [24]

    Differential evolution–a simple and effi- cient heuristic for global optimization over continuous spaces,

    R. Storn and K. Price, “Differential evolution–a simple and effi- cient heuristic for global optimization over continuous spaces,” Journal of global optimization , vol. 11, pp. 341–359, 1997

  17. [25]

    Back, Evolutionary algorithms in theory and practice: evolu- tion strategies, evolutionary programming, genetic algorithms

    T. Back, Evolutionary algorithms in theory and practice: evolu- tion strategies, evolutionary programming, genetic algorithms . Oxford University Press, 1996

  18. [26]

    Self-adaptive differential evolution algo- rithm with zoning evolution of control parameters and adaptive mutation strategies,

    Q. Fan and X. Yan, “Self-adaptive differential evolution algo- rithm with zoning evolution of control parameters and adaptive mutation strategies,” IEEE transactions on cybernetics, vol. 46, no. 1, pp. 219–232, 2015

  19. [27]

    Reinforcement learning: An introduction,

    R. S. Sutton, “Reinforcement learning: An introduction,” A Bradford Book, 2018

  20. [28]

    Q-learning: Theory and applications,

    J. Clifton and E. Laber, “Q-learning: Theory and applications,” Annual Review of Statistics and Its Application , vol. 7, no. 1, pp. 279–301, 2020

  21. [29]

    G. A. Rummery and M. Niranjan, On-line Q-learning using connectionist systems. University of Cambridge, Department of Engineering Cambridge, UK, 1994, vol. 37

  22. [30]

    Pol- icy gradient methods for reinforcement learning with function approximation,

    R. S. Sutton, D. McAllester, S. Singh, and Y. Mansour, “Pol- icy gradient methods for reinforcement learning with function approximation,” Advances in neural information processing sys- tems, vol. 12, 1999

  23. [31]

    A sur- vey of actor-critic reinforcement learning: Standard and natural policy gradients,

    I. Grondman, L. Busoniu, G. A. Lopes, and R. Babuska, “A sur- vey of actor-critic reinforcement learning: Standard and natural policy gradients,” IEEE Transactions on Systems, Man, and Cybernetics, part C (applications and reviews) , vol. 42, no. 6, pp. 1291–1307, 2012

  24. [32]

    Reinforced model predictive control (rl-mpc) for building energy manage- ment,

    J. Arroyo, C. Manna, F. Spiessens, and L. Helsen, “Reinforced model predictive control (rl-mpc) for building energy manage- ment,” Applied Energy, vol. 309, p. 118346, 2022

  25. [33]

    Safe reinforcement learning for autonomous vehicle using monte carlo tree search,

    S. Mo, X. Pei, and C. Wu, “Safe reinforcement learning for autonomous vehicle using monte carlo tree search,”IEEE Trans- actions on Intelligent Transportation Systems, vol. 23, no. 7, pp. 6766–6773, 2021

  26. [34]

    Reinforce- ment learning-assisted evolutionary algorithm: A survey and research opportunities,

    Y. Song, Y. Wu, Y. Guo, R. Yan, P. N. Suganthan, Y. Zhang, W. Pedrycz, S. Das, R. Mallipeddi, O. S. Ajaniet al., “Reinforce- ment learning-assisted evolutionary algorithm: A survey and research opportunities,” Swarm and Evolutionary Computation, vol. 86, p. 101517, 2024

  27. [35]

    Bridging evolutionary algorithms and reinforcement learning: A compre- hensive survey on hybrid algorithms,

    P. Li, J. Hao, H. Tang, X. Fu, Y. Zhen, and K. Tang, “Bridging evolutionary algorithms and reinforcement learning: A compre- hensive survey on hybrid algorithms,” IEEE Transactions on Evolutionary Computation, 2024

  28. [36]

    Deep reinforcement learning based parameter control in differ- ential evolution,

    M. Sharma, A. Komninos, M. L´ opez-Ib´ a˜ nez, and D. Kazakov, “Deep reinforcement learning based parameter control in differ- ential evolution,” in Proceedings of the genetic and evolutionary computation conference, 2019, pp. 709–717

  29. [37]

    Controlling sequential hybrid evolutionary algorithm by q-learning [research frontier][research frontier],

    H. Zhang, J. Sun, T. B¨ ack, Q. Zhang, and Z. Xu, “Controlling sequential hybrid evolutionary algorithm by q-learning [research frontier][research frontier],” IEEE Computational Intelligence Magazine, vol. 18, no. 1, pp. 84–103, 2023

  30. [38]

    Reinforcement learning-based differential evolution for parameters extraction of photovoltaic models,

    Z. Hu, W. Gong, and S. Li, “Reinforcement learning-based differential evolution for parameters extraction of photovoltaic models,” Energy Reports, vol. 7, pp. 916–928, 2021

  31. [39]

    Understanding the problem space in single-objective numerical optimization us- ing exploratory landscape analysis,

    U. ˇSkvorc, T. Eftimov, and P. Koroˇ sec, “Understanding the problem space in single-objective numerical optimization us- ing exploratory landscape analysis,” Applied Soft Computing , vol. 90, p. 106138, 2020

  32. [40]

    flaccogui: Exploratory landscape analysis for everyone,

    C. Hanster and P. Kerschke, “flaccogui: Exploratory landscape analysis for everyone,” in Proceedings of the Genetic and Evolu- tionary Computation Conference Companion , 2017, pp. 1215– 1222

  33. [41]

    Towards explain- able exploratory landscape analysis: extreme feature selection for classifying bbob functions,

    Q. Renau, J. Dreo, C. Doerr, and B. Doerr, “Towards explain- able exploratory landscape analysis: extreme feature selection for classifying bbob functions,” in Applications of Evolutionary Computation: 24th International Conference, EvoApplications 2021, Held as Part of EvoSta...

  34. [42]

    Au- tomated algorithm selection: Survey and perspectives,

    P. Kerschke, H. H. Hoos, F. Neumann, and H. Trautmann, “Au- tomated algorithm selection: Survey and perspectives,” Evolu- tionary computation, vol. 27, no. 1, pp. 3–45, 2019. IEEE, VOL. X, NO. X, X X 12

  35. [43]

    Exploratory landscape anal- ysis for mixed-variable problems,

    R. P. Prager and H. Trautmann, “Exploratory landscape anal- ysis for mixed-variable problems,” IEEE Transactions on Evo- lutionary Computation, 2024

  36. [44]

    New features for continuous exploratory landscape anal- ysis based on the soo tree,

    B. Derbel, A. Liefooghe, S. V´ erel, H. Aguirre, and K. Tanaka, “New features for continuous exploratory landscape anal- ysis based on the soo tree,” in Proceedings of the 15th ACM/SIGEVO Conference on Foundations of Genetic Algo- rithms, 2019, pp. 72–86

  37. [45]

    Cell mapping techniques for exploratory landscape analysis,

    P. Kerschke, M. Preuss, C. Hern´ andez, O. Sch¨ utze, J.-Q. Sun, C. Grimme, G. Rudolph, B. Bischl, and H. Trautmann, “Cell mapping techniques for exploratory landscape analysis,” in EVOLVE-A Bridge between Probability, Set Oriented Numerics, and Evolutionary Computation V. Spr...

  38. [46]

    Exploratory landscape analysis of continuous space optimization problems using information content,

    M. A. Mu˜ noz, M. Kirley, and S. K. Halgamuge, “Exploratory landscape analysis of continuous space optimization problems using information content,” IEEE transactions on evolutionary computation, vol. 19, no. 1, pp. 74–87, 2014

  39. [47]

    Pflacco: Feature-based landscape analysis of continuous and constrained optimization problems in python,

    R. Prager and H. Trautmann, “Pflacco: Feature-based landscape analysis of continuous and constrained optimization problems in python,” Evolutionary Computation , vol. 32, no. 3, pp. 211–216, 09 2024. [Online]. Available: https://doi.org/10.1162/ evco a 00341

  40. [48]

    Expressiveness and robustness of landscape features,

    Q. Renau, J. Dreo, C. Doerr, and B. Doerr, “Expressiveness and robustness of landscape features,” in Proceedings of the Genetic and Evolutionary Computation Conference Companion , 2019, pp. 2048–2051

  41. [49]

    De- tecting funnel structures by means of exploratory landscape analysis,

    P. Kerschke, M. Preuss, S. Wessing, and H. Trautmann, “De- tecting funnel structures by means of exploratory landscape analysis,” in Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation , 2015, pp. 265–272

  42. [50]

    The dispersion metric and the cma evolution strategy,

    M. Lunacek and D. Whitley, “The dispersion metric and the cma evolution strategy,” in Proceedings of the 8th annual conference on Genetic and evolutionary computation , 2006, pp. 477–484

  43. [51]

    Exploratory land- scape analysis is strongly sensitive to the sampling strategy,

    Q. Renau, C. Doerr, J. Dreo, and B. Doerr, “Exploratory land- scape analysis is strongly sensitive to the sampling strategy,” in Parallel Problem Solving from Nature–PPSN XVI: 16th In- ternational Conference, PPSN 2020, Leiden, The Netherlands, September 5-9, 2020, Proceedings...

  44. [52]

    Low- budget exploratory landscape analysis on multiple peaks mod- els,

    P. Kerschke, M. Preuss, S. Wessing, and H. Trautmann, “Low- budget exploratory landscape analysis on multiple peaks mod- els,” in Proceedings of the Genetic and Evolutionary Computa- tion Conference 2016, 2016, pp. 229–236

  45. [53]

    Differential evolution for economic load dispatch problems,

    N. Noman and H. Iba, “Differential evolution for economic load dispatch problems,” Electric power systems research , vol. 78, no. 8, pp. 1322–1331, 2008

  46. [54]

    Review of differential evolution population size,

    A. P. Piotrowski, “Review of differential evolution population size,” Swarm and Evolutionary Computation , vol. 32, pp. 1–24, 2017

  47. [55]

    An efficient constraint handling method for genetic algorithms,

    K. Deb, “An efficient constraint handling method for genetic algorithms,” Computer methods in applied mechanics and engi- neering, vol. 186, no. 2-4, pp. 311–338, 2000

  48. [56]

    Array programming with NumPy,

    C. R. Harris, K. J. Millman, S. J. van der Walt, R. Gommers, P. Virtanen, D. Cournapeau, E. Wieser, J. Taylor, S. Berg, N. J. Smith, R. Kern, M. Picus, S. Hoyer, M. H. van Kerkwijk, M. Brett, A. Haldane, J. F. del R´ ıo, M. Wiebe, P. Peterson, P. G´ erard-Marchant, K. Sheppard...

  49. [57]

    Introduction to multi-armed bandits,

    A. Slivkins et al., “Introduction to multi-armed bandits,” Foun- dations and Trends® in Machine Learning, vol. 12, no. 1-2, pp. 1–286, 2019

  50. [58]

    Learning to utilize shaping rewards: A new approach of reward shaping,

    Y. Hu, W. Wang, H. Jia, Y. Wang, Y. Chen, J. Hao, F. Wu, and C. Fan, “Learning to utilize shaping rewards: A new approach of reward shaping,” Advances in Neural Information Processing Systems, vol. 33, pp. 15 931–15 941, 2020

  51. [59]

    Real-parameter black-box optimization benchmarking 2009: Noiseless functions definitions,

    N. Hansen, S. Finck, R. Ros, and A. Auger, “Real-parameter black-box optimization benchmarking 2009: Noiseless functions definitions,” Ph.D. dissertation, INRIA, 2009

  52. [60]

    Metabox: a benchmark platform for meta-black-box optimization with reinforcement learning,

    Z. Ma, H. Guo, J. Chen, Z. Li, G. Peng, Y.-J. Gong, Y. Ma, and Z. Cao, “Metabox: a benchmark platform for meta-black-box optimization with reinforcement learning,” Advances in Neural Information Processing Systems, vol. 36, 2024

  53. [61]

    Differential evolution with hybrid parameters and mutation strategies based on reinforcement learning,

    Z. Tan, Y. Tang, K. Li, H. Huang, and S. Luo, “Differential evolution with hybrid parameters and mutation strategies based on reinforcement learning,” Swarm and Evolutionary Computation , vol. 75, p. 101194, 2022. [Online]. Available: https://www.sciencedirect.com/science/arti...

  54. [62]

    Self-adaptive differen- tial evolution algorithm with population size reduction for single objective bound-constrained optimization: Algorithm j21,

    J. Brest, M. S. Mauˇ cec, and B. Boˇ skovi´ c, “Self-adaptive differen- tial evolution algorithm with population size reduction for single objective bound-constrained optimization: Algorithm j21,” in 2021 IEEE Congress on Evolutionary Computation (CEC) . IEEE, 2021, pp. 817–824

  55. [63]

    Improving differential evolution through bayesian hyperpa- rameter optimization,

    S. Biswas, D. Saha, S. De, A. D. Cobb, S. Das, and B. A. Jalaian, “Improving differential evolution through bayesian hyperpa- rameter optimization,” in 2021 IEEE congress on evolutionary computation (CEC). IEEE, 2021, pp. 832–840

  56. [64]

    Nl-shade-lbc algorithm with linear parameter adaptation bias change for cec 2022 numerical optimization,

    V. Stanovov, S. Akhmedova, and E. Semenkin, “Nl-shade-lbc algorithm with linear parameter adaptation bias change for cec 2022 numerical optimization,” in 2022 IEEE Congress on Evolutionary Computation (CEC) . IEEE, 2022, pp. 01–08. Xu Y ang is a PhD student majoring in man- ag...

Pith tools

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