pith. machine review for the scientific record. sign in

arxiv: 2602.16044 · v2 · submitted 2026-02-17 · ⚛️ physics.comp-ph

Recognition: 2 theorem links

· Lean Theorem

Multi-Objective Evolutionary Design of Molecules with Enhanced Nonlinear Optical Properties

Dominic Mashak , Jacob Schrum , S.A. Alexander
Authors on Pith no claims yet

Pith reviewed 2026-05-15 21:38 UTC · model grok-4.3

classification ⚛️ physics.comp-ph
keywords nonlinear opticsmolecular designevolutionary algorithmsNSGA-IIquality diversityhyperpolarizabilitySMILESmulti-objective optimization
0
0 comments X

The pith

Evolutionary algorithms design molecules optimizing nonlinear optical properties across four objectives.

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

The paper compares several evolutionary algorithms for discovering molecules with improved nonlinear optical properties. It targets maximizing the ratio of first to second hyperpolarizability while tuning the HOMO-LUMO gap and polarizability to useful ranges and minimizing energy. Quality diversity approaches like MOME explore more of the chemical space than NSGA-II, achieving higher overall performance metrics despite NSGA-II finding molecules strong in individual objectives. This matters because better NLO materials could advance photonics and laser technologies by providing molecules that balance competing physical requirements.

Core claim

By encoding molecules as SMILES strings and evaluating them with quantum-chemical calculations, the study shows that NSGA-II produces molecules with consistently high scores in each objective, while MOME achieves superior global hypervolume and MOQD scores by maintaining a diverse archive across atom and bond count measures.

What carries the argument

Quality-diversity evolutionary algorithms such as MOME and MAP-Elites that maintain archives in a measure space defined by molecular size and bond count, combined with multi-objective optimization of hyperpolarizability ratio, orbital gap, polarizability, and energy.

Load-bearing premise

The quantum-chemical calculations used to score the molecules accurately predict their actual nonlinear optical behavior in real materials.

What would settle it

Synthesizing the highest-scoring molecules from the evolutionary runs and experimentally measuring their hyperpolarizabilities and other properties to compare against the computed values.

Figures

Figures reproduced from arXiv: 2602.16044 by Dominic Mashak, Jacob Schrum, S.A. Alexander.

Figure 1
Figure 1. Figure 1: Median Best Objective Scores Across 20 Runs of Each Algorithm: (a) Median first-to-second hyperpolarizability ratio. High values are better, so (𝜇 + 𝜆) outperforms all others by a large margin, including other single-objective methods, though NSGA-II is clearly second-best. (b) Median linear polarizability range deviation. All but simulated annealing and (𝜇 + 𝜆) quickly reach a perfect minimal score of 0, … view at source ↗
Figure 2
Figure 2. Figure 2: Global Hypervolume Scores Across 20 Runs of Each Algorithm: (a) Median hypervolume scores for each algorithm across function evaluations. MOME𝐹 is the best, followed by a cluster of (𝜇 + 𝜆), NSGA-II, and MAP-Elites𝐶 , before algorithms start to bunch together near the bottom. (b) Box-and-whisker plots of hypervolume scores for final Pareto fronts. The lower quartile, median, and upper quartile are the lowe… view at source ↗
Figure 3
Figure 3. Figure 3: Fine-grained and Coarse Archive Median Scores Across 20 Runs of Each Algorithm: (a) Median bin count with fine-grained binning. Unsurprisingly, QD methods fill more bins than non-QD approaches, and simulated annealing performs the worst. Fine-grained QD methods occupy slightly more bins than their coarse counterparts. (b) Median bin count with coarse binning. The scale is different with a coarse archive, b… view at source ↗
Figure 4
Figure 4. Figure 4: Fine-grained Archive Median QD Scores by Objective Across 20 Runs of Each Algorithm: (a) Median QD for first-to-second hyperpolarizability ratio using fine-grained binning. Qualitatively similar to raw objective scores for 𝛽/𝛾 (Figure 4a), with strong performance by (𝜇 + 𝜆) and NSGA-II in second place. (b) Median QD for linear polarizability range deviation using fine-grained binning. Both MOME approaches … view at source ↗
Figure 5
Figure 5. Figure 5: Fine-grained Mega Archive Hypervolume Heatmaps: Fine-grained archives that combine solutions from each algorithm across all 20 seeds, with heat scale showing each bin’s HV score. The x-axis is the atom count and the y-axis is the bond count. 5 10 15 20 25 30 Number of Atoms 5 10 15 20 25 30 Number of Bonds 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Hypervolume (a) SA 5 10 15 20 25 30 Number of Atoms 5 10 15 20 25 30 Numb… view at source ↗
Figure 6
Figure 6. Figure 6: Coarse Mega Archive Hypervolume Heatmaps: Coarse archives that combine solutions from each algorithm across all 20 seeds, with heat scale showing each bin’s HV score. The x-axis is the atom count and the y-axis is the bond count. neglecting secondary objectives. Notably, (𝜇 + 𝜆) solutions showed awful 𝑓𝛼 compared to others that achieved perfect targeting of 0 a.u. The large 𝑓𝛼 values indicate molecules tha… view at source ↗
read the original abstract

Nonlinear optical (NLO) materials are essential for many photonic, telecommunication, and laser technologies, yet discovering better NLO molecules is computationally challenging due to the vast chemical space and competing objectives. We compare evolutionary algorithms for molecular design, targeting four objectives: maximizing the ratio of first-to-second hyperpolarizability $(\beta/\gamma)$, optimizing HOMO-LUMO gap and linear polarizability to target ranges, and minimizing energy per atom. We encode molecules as SMILES strings and evaluate their properties using quantum-chemical calculations. We compare NSGA-II, MAP-Elites, MOME, a single-objective $(\mu+\lambda)$ evolutionary algorithm, and simulated annealing. Quality diversity methods maintain archives across a measure space defined by atom and bond count, enabling the discovery of structurally diverse molecules. Our results demonstrate that NSGA-II consistently earns high scores in every objective, leading to high-quality molecules, but MOME does a better job exploring a wide range of possibilities, resulting in higher global hypervolume and MOQD scores. However, each method has strengths and weaknesses, and produced many promising molecules.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper compares evolutionary algorithms (NSGA-II, MAP-Elites, MOME, single-objective EA, simulated annealing) for multi-objective molecular design targeting enhanced nonlinear optical properties. Molecules are represented as SMILES strings and evaluated via quantum-chemical calculations on four objectives: maximizing the β/γ ratio, optimizing HOMO-LUMO gap and linear polarizability to target ranges, and minimizing energy per atom. Quality-diversity methods maintain archives over a measure space of atom and bond counts. The central claim is that NSGA-II yields high per-objective scores while MOME achieves superior global hypervolume and MOQD through broader exploration, with all methods producing promising molecules.

Significance. If the quantum-chemical oracles are reliable, the work provides empirical evidence that archive-based quality-diversity algorithms can discover structurally diverse molecules with competitive NLO performance, which is relevant for photonic materials discovery. The direct comparison of NSGA-II against MOME on hypervolume/MOQD metrics offers a concrete benchmark for multi-objective molecular optimization.

major comments (3)
  1. [Methods] Methods section on property evaluation: the specific quantum-chemical level of theory (DFT functional, basis set, solvation model) used to compute β and γ is not stated. Hyperpolarizabilities are known to vary strongly with these choices; without this detail the reported performance gaps between MOME and NSGA-II cannot be verified as genuine algorithmic differences rather than artifacts of the oracle.
  2. [Results] Results section, hypervolume and MOQD comparisons: no error bars, standard deviations, or statistical significance tests from repeated runs are provided for the global hypervolume and MOQD scores. This makes it impossible to assess whether MOME's reported advantage is robust or attributable to stochastic variation in the evolutionary runs.
  3. [Results] Results section, objective definitions: the target ranges for HOMO-LUMO gap and polarizability are presented without reference to experimental benchmarks or justification that they correspond to practically useful NLO materials, weakening the claim that the discovered molecules are 'promising'.
minor comments (2)
  1. [Abstract] Abstract: the phrase 'many promising molecules' is used without a quantitative definition tied to the four objectives or a count of molecules meeting all thresholds.
  2. [Figures] Figure captions: the measure-space plots for atom/bond count archives lack axis labels specifying the exact discretization bins used by MAP-Elites and MOME.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which have helped us identify areas for improvement in clarity and rigor. We address each major comment point-by-point below and will revise the manuscript accordingly to enhance reproducibility and strengthen the claims.

read point-by-point responses

  1. Referee: [Methods] Methods section on property evaluation: the specific quantum-chemical level of theory (DFT functional, basis set, solvation model) used to compute β and γ is not stated. Hyperpolarizabilities are known to vary strongly with these choices; without this detail the reported performance gaps between MOME and NSGA-II cannot be verified as genuine algorithmic differences rather than artifacts of the oracle.

    Authors: We agree that the absence of these computational details limits reproducibility and could confound interpretation of algorithmic differences. The calculations were performed using DFT at the B3LYP/6-31G(d) level with the default gas-phase model in the underlying quantum chemistry package; we will add a dedicated subsection in the Methods detailing the exact functional, basis set, and any solvation settings (none were used) along with software versions and convergence criteria. This revision will allow readers to confirm that performance differences arise from the optimization algorithms rather than oracle variations. revision: yes

  2. Referee: [Results] Results section, hypervolume and MOQD comparisons: no error bars, standard deviations, or statistical significance tests from repeated runs are provided for the global hypervolume and MOQD scores. This makes it impossible to assess whether MOME's reported advantage is robust or attributable to stochastic variation in the evolutionary runs.

    Authors: We acknowledge that single-run results do not adequately demonstrate robustness against stochasticity. In the revised manuscript we will report aggregated statistics from 10 independent runs with different random seeds, including mean hypervolume and MOQD values with standard deviation error bars. We will also add Wilcoxon rank-sum tests (with p-values) comparing MOME against NSGA-II and the other baselines to establish statistical significance of the observed advantages. revision: yes

  3. Referee: [Results] Results section, objective definitions: the target ranges for HOMO-LUMO gap and polarizability are presented without reference to experimental benchmarks or justification that they correspond to practically useful NLO materials, weakening the claim that the discovered molecules are 'promising'.

    Authors: The target ranges were selected to align with values commonly associated with organic NLO chromophores in the literature (HOMO-LUMO gaps of 2.0–4.0 eV for visible/near-IR transparency and polarizabilities of 150–600 a.u. for enhanced response in push-pull systems). We will expand the manuscript to include explicit citations to experimental and computational benchmarks (e.g., studies on stilbene and thiophene derivatives) that justify these intervals as practically relevant for photonic applications, thereby reinforcing the relevance of the discovered molecules. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical algorithm comparison against external quantum oracles

full rationale

The manuscript is an empirical benchmarking study that applies standard multi-objective evolutionary algorithms (NSGA-II, MOME, MAP-Elites, etc.) to a molecular design task. Fitness values for the four objectives (β/γ ratio, HOMO-LUMO gap, polarizability, energy/atom) are obtained from independent quantum-chemical calculations performed on SMILES-encoded structures. No derivations, parameter fits, or uniqueness theorems are claimed; the central results are direct performance metrics (hypervolume, MOQD) computed from those external oracle evaluations. No self-citations are load-bearing, no ansatzes are smuggled, and no quantities are redefined in terms of themselves. The work is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the assumption that the four chosen objectives are sufficient proxies for useful NLO performance and that the quantum-chemistry evaluator is reliable; no new entities are postulated.

axioms (1)
  • domain assumption Quantum-chemical calculations (unspecified level of theory) provide accurate enough values for β, γ, HOMO-LUMO gap, polarizability, and energy to guide evolutionary search.
    Invoked when the paper states that properties are evaluated using quantum-chemical calculations.

pith-pipeline@v0.9.0 · 5488 in / 1300 out tokens · 15702 ms · 2026-05-15T21:38:58.804268+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. CVT Archives and Chemical Embedding Measures for Multi-Objective Quality Diversity in Molecular Design

    physics.comp-ph 2026-04 unverdicted novelty 5.0

    CVT archives with learned chemical embeddings improve median global hypervolume and multi-objective quality diversity in NLO molecular design compared to grid-based archives.

Reference graph

Works this paper leans on

48 extracted references · 48 canonical work pages · cited by 1 Pith paper · 3 internal anchors

  1. [1]

    Hans-Georg Beyer and Hans-Paul Schwefel. 2002. Evolution strategies - A comprehensive introduction.Nat. Comput.1, 1 (2002), 3–52. http://dblp.uni- trier.de/db/journals/nc/nc1.html#BeyerS02

    work page 2002

  2. [2]

    David M. Bishop. 1998.Molecular Vibration and Nonlinear Optics. John Wiley & Sons, Ltd, 1–40. arXiv:https://onlinelibrary.wiley.com/doi/pdf/10.1002/9780470141632.ch1 doi:10.1002/9780470141632.ch1 8

    work page doi:10.1002/9780470141632.ch1 1998

  3. [3]

    Blaney and J

    Jeffrey M. Blaney and J. Scott Dixon. 1991. Chapter 29. Receptor Modeling by Distance Geometry.Annual Reports in Medicinal Chemistry26 (1991), 281–285. https://api.semanticscholar.org/CorpusID:94219120

    work page 1991

  4. [4]

    Blank and K

    J. Blank and K. Deb. 2020. pymoo: Multi-Objective Optimization in Python.IEEE Access8 (2020), 89497–89509

    work page 2020

  5. [5]

    Antoine Cully, Jeff Clune, Danesh Tarapore, and Jean-Baptiste Mouret. 2015. Robots that can adapt like animals.Nature521, 7553 (2015), 503–507. doi:10.1038/ NATURE14422

    work page 2015

  6. [6]

    Larry Dalton, Bruce Robinson, Alex Jen, Philip Ried, Bruce Eichinger, Sei-Hum Jang, Jingdong Luo, Sen Liu, Yi Liao, Kimberly Firestone, Nishant Bhatambrekar, and Denise Bale. 2004. Organic electro-optic materials.Proceedings of SPIE - The International Society for Optical Engineering5621 (12 2004). doi:10.1117/12.584102

    work page doi:10.1117/12.584102 2004

  7. [7]

    Dastmalchi, A

    P. Dastmalchi, A. Haddadpour, and G. Veronis. 2014. 11 - Nanophotonics: devices for manipulating light at the nanoscale. InNanolithography, Martin Feldman (Ed.). Woodhead Publishing, 376–398. doi:10.1533/9780857098757.376

    work page doi:10.1533/9780857098757.376 2014

  8. [8]

    K. Deb, A. Pratap, S. Agarwal, and T. Meyarivan. 2002. A fast and elitist multiobjec- tive genetic algorithm: NSGA-II.IEEE Transactions on Evolutionary Computation 6, 2 (2002), 182–197. doi:10.1109/4235.996017

    work page doi:10.1109/4235.996017 2002

  9. [9]

    Dowsland

    Kathryn A. Dowsland. 1993.Modern Heuristic Techniques for Combinatorial Problems. John Wiley & Sons, Inc., USA, 20–69

    work page 1993

  10. [10]

    Fontaine and Stefanos Nikolaidis

    Matthew C. Fontaine and Stefanos Nikolaidis. 2021. Differentiable quality di- versity. InProceedings of the 35th International Conference on Neural Information Processing Systems (NIPS ’21). Curran Associates Inc., Red Hook, NY, USA, Article 768, 13 pages

    work page 2021

  11. [11]

    Thomas A. Halgren. 1996. Merck molecular force field. I. Basis, form, scope, parameterization, and performance of MMFF94.Journal of Computational Chem- istry17, 5-6 (1996), 490–519. doi:10.1002/(SICI)1096-987X(199604)17:5/6<490:: AID-JCC1>3.0.CO;2-P

    work page doi:10.1002/(sici)1096-987x(199604)17:5/6 1996

  12. [12]

    Tomoya Hömberg, Sanaz Mostaghim, Satoru Hiwa, and Tomoyuki Hiroyasu

    work page

  13. [13]

    Critical current of a Josephson junction containing a conical magnet

    Optimized Drug Design using Multi-Objective Evolutionary Algorithms with SELFIES. In2024 IEEE Congress on Evolutionary Computation (CEC). 1–8. doi:10.1109/CEC60901.2024.10611870

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1109/cec60901.2024.10611870 2024

  14. [14]

    Tomáš Hrivnák, Miroslav Medveď, Wojciech Bartkowiak, and Robert Zaleśny

    work page

  15. [15]

    Hyperpolarizabilities of Push–Pull Chromophores in Solution: Interplay between Electronic and Vibrational Contributions.Molecules27, 24 (2022). doi:10. 3390/molecules27248738

    work page 2022

  16. [16]

    Kanis, Mark A

    David R. Kanis, Mark A. Ratner, and Tobin J. Marks. 1994. Design and con- struction of molecular assemblies with large second-order optical nonlinear- ities. Quantum chemical aspects.Chemical Reviews94, 1 (1994), 195–242. arXiv:https://doi.org/10.1021/cr00025a007 doi:10.1021/cr00025a007

    work page doi:10.1021/cr00025a007 1994

  17. [17]

    Kurtz, James J

    Henry A. Kurtz, James J. P. Stewart, and Kenneth M. Di- eter. 1990. Calculation of the nonlinear optical properties of molecules.Journal of Computational Chemistry11, 1 (1990), 82–

    work page 1990

  18. [18]

    arXiv:https://onlinelibrary.wiley.com/doi/pdf/10.1002/jcc.540110110 doi:10.1002/jcc.540110110

    work page doi:10.1002/jcc.540110110

  19. [19]

    M.G. Kuzyk. 2001. Quantum limits of the hyper-Rayleigh scattering susceptibility. IEEE Journal of Selected Topics in Quantum Electronics7, 5 (2001), 774–780. doi:10. 1109/2944.979338

    work page arXiv 2001

  20. [20]

    Mark G. Kuzyk. 2000. Fundamental limits on third-order molecular susceptibili- ties.Opt. Lett.25, 16 (Aug 2000), 1183–1185. doi:10.1364/OL.25.001183

    work page doi:10.1364/ol.25.001183 2000

  21. [21]

    Mark G. Kuzyk. 2000. Physical Limits on Electronic Nonlinear Molecular Sus- ceptibilities.Phys. Rev. Lett.85 (Aug 2000), 1218–1221. Issue 6. doi:10.1103/ PhysRevLett.85.1218

    work page 2000

  22. [22]

    2010.RDKit: Open-source cheminformatics

    Greg Landrum. 2010.RDKit: Open-source cheminformatics. https://www.rdkit.org

    work page 2010

  23. [23]

    S. R. Marder, D. N. Beratan, and L.-T. Cheng. 1991. Approaches for Optimizing the First Electronic Hyperpolarizability of Conjugated Organic Molecules.Science 252, 5002 (1991), 103–106. http://www.jstor.org/stable/2875581

    work page arXiv 1991

  24. [24]

    Seth R Marder, Bernard Kippelen, Alex K-Y Jen, and Nasser Peyghambarian

    work page

  25. [25]

    doi:10.1038/42190

    Design and synthesis of chromophores and polymers for electro-optic and photorefractive applications.Nature388, 6645 (1997), 845–851. doi:10.1038/42190

    work page doi:10.1038/42190 1997

  26. [26]

    Dominic Mashak and Steven Alexander. 2025. Finding Molecules with Large Hy- perpolarizabilities. InMATCH Commun. Math. Comput. Chem., Vol. 94. MATCH Commun. Math. Comput. Chem., 633–644. Issue 3. doi:10.46793/match94-3.25824

    work page doi:10.46793/match94-3.25824 2025

  27. [27]

    Dominic Mashak and Steven Alexander. 2025. Finding Molecules with Specific Properties: Simulated Annealing vs. Evolution. InProceedings of the Genetic and Evolutionary Computation Conference Companion(NH Malaga Hotel, Malaga, Spain)(GECCO ’25 Companion). Association for Computing Machinery, New York, NY, USA, 759–762. doi:10.1145/3712255.3726635

    work page doi:10.1145/3712255.3726635 2025

  28. [28]

    Dominic Mashak and S. A. Alexander. 2025. Benchmarking Hartree-Fock and DFT for Molecular Hyperpolarizability: Implications for Evolutionary Design. arXiv:2511.17767 [physics.chem-ph] https://arxiv.org/abs/2511.17767

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  29. [29]

    Rosenbluth, Marshall N

    Nicholas Metropolis, Arianna W. Rosenbluth, Marshall N. Rosenbluth, Augusta H. Teller, and Edward Teller. 1953. Equation of State Calculations by Fast Computing Machines.The Journal of Chemical Physics21, 6 (06 1953), 1087–1092. doi:10. 1063/1.1699114

    work page 1953

  30. [30]

    Minasian

    R.A. Minasian. 2005. Modulation and Demodulation of Optical Signals. InEncy- clopedia of Modern Optics, Bob D. Guenther and Duncan G. Steel (Eds.). Elsevier, Oxford, 129–138. https://www.sciencedirect.com/referencework/9780123693952/ encyclopedia-of-modern-optics

    work page arXiv 2005

  31. [31]

    Jean-Baptiste Mouret and Jeff Clune. 2015. Illuminating search spaces by mapping elites. arXiv:1504.04909 [cs.AI] https://arxiv.org/abs/1504.04909

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  32. [32]

    Lucjan Piela. 2020. The Molecule Subject to Electric or Magnetic Fields. InIdeas of Quantum Chemistry (Third Edition)(third edition ed.), Lucjan Piela (Ed.). Elsevier, 253–335. doi:10.1016/B978-0-44-464248-6.00012-0

    work page doi:10.1016/b978-0-44-464248-6.00012-0 2020

  33. [33]

    Thomas Pierrot, Guillaume Richard, Karim Beguir, and Antoine Cully. 2022. Multi-objective quality diversity optimization. InProceedings of the Genetic and Evolutionary Computation Conference(Boston, Massachusetts)(GECCO ’22). Association for Computing Machinery, New York, NY, USA, 139–147. doi:10.1145/3512290.3528823

    work page doi:10.1145/3512290.3528823 2022

  34. [34]

    Justin K. Pugh, L. B. Soros, Paul A. Szerlip, and Kenneth O. Stanley. 2015. Confronting the Challenge of Quality Diversity. InProceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation(Madrid, Spain) (GECCO ’15). Association for Computing Machinery, New York, NY, USA, 967–974. doi:10.1145/2739480.2754664

    work page doi:10.1145/2739480.2754664 2015

  35. [35]

    Kuzyk, Juefei Zhou, Shiva K

    Javier Pérez-Moreno, Sheng-Ting Hung, Mark G. Kuzyk, Juefei Zhou, Shiva K. Ramini, and Koen Clays. 2011. Experimental verification of a self-consistent theory of the first-, second-, and third-order (non)linear optical response.Physical Review A84, 3 (Sept. 2011). doi:10.1103/physreva.84.033837

    work page doi:10.1103/physreva.84.033837 2011

  36. [36]

    Rappé, Carla Casewit, K

    Anthony K. Rappé, Carla Casewit, K. S. Colwell, William A. Goddard, and W. M. Skiff. 1992. UFF, a full periodic table force field for molecular mechanics and molecular dynamics simulations.Journal of the American Chemical Society114 (1992), 10024–10035. https://api.semanticscholar.org/CorpusID:17479693

    work page 1992

  37. [37]

    Blanca Ros

    M. Blanca Ros. 2008. Organic Materials for Nonlinear Optics. InEngineering of Crystalline Materials Properties, Juan J. Novoa, Dario Braga, and Lia Addadi (Eds.). Springer Netherlands, Dordrecht, 375–390

    work page 2008

  38. [38]

    Bahaa E. A. Saleh and Malvin Carl Teich. 1991.Electro- Optics. John Wiley & Sons, Ltd, NY, Chapter 18, 696–736. arXiv:https://onlinelibrary.wiley.com/doi/pdf/10.1002/0471213748.ch18 doi:10.1002/0471213748.ch18

    work page doi:10.1002/0471213748.ch18 1991

  39. [39]

    K. D. Singer, J. E. Sohn, L. A. King, H. M. Gordon, H. E. Katz, and C. W. Dirk. 1989. Second-order nonlinear-optical properties of donor- and acceptor-substituted aromatic compounds.J. Opt. Soc. Am. B6, 7 (Jul 1989), 1339–1350. doi:10.1364/ JOSAB.6.001339

    work page 1989

  40. [40]

    Sullivan and Larry R

    Philip A. Sullivan and Larry R. Dalton. 2010. Theory-Inspired Development of Organic Electro-optic Materials.Accounts of Chemical Research43, 1 (2010), 10–18. arXiv:https://doi.org/10.1021/ar800264w doi:10.1021/ar800264w PMID: 19663413

    work page doi:10.1021/ar800264w 2010

  41. [41]

    Blunt, Nikolay A

    Qiming Sun, Xing Zhang, Samragni Banerjee, Peng Bao, Marc Barbry, Nick S. Blunt, Nikolay A. Bogdanov, George H. Booth, Jia Chen, Zhi-Hao Cui, Janus J. Eriksen, Yang Gao, Sheng Guo, Jan Hermann, Matthew R. Hermes, Kevin Koh, Peter Koval, Susi Lehtola, Zhendong Li, Junzi Liu, Narbe Mardirossian, James D. McClain, Mario Motta, Bastien Mussard, Hung Q. Pham, ...

    work page doi:10.1063/5.0006074 2020

  42. [42]

    Attila Szabó and Neil S. Ostlund. 1996. Modern quantum chemistry : introduc- tion to advanced electronic structure theory. https://api.semanticscholar.org/ CorpusID:94743139

    work page 1996

  43. [43]

    Peter J. M. van Laarhoven and Emile H. L. Aarts. 1987. Simulated Annealing: Theory and Applications. InMathematics and Its Applications. Springer. https: //link.springer.com/book/10.1007/978-94-015-7744-1

    work page doi:10.1007/978-94-015-7744-1 1987

  44. [44]

    Jonas Verhellen. 2022. Graph-based molecular Pareto optimisation.Chem. Sci.13 (2022), 7526–7535. Issue 25. doi:10.1039/D2SC00821A

    work page doi:10.1039/d2sc00821a 2022

  45. [45]

    Jonas Verhellen and Jeriek Van den Abeele. 2020. Illuminating elite patches of chemical space.Chem. Sci.11 (2020), 11485–11491. Issue 42. doi:10.1039/ D0SC03544K

    work page 2020

  46. [46]

    David Weininger. 1988. SMILES, a chemical language and informa- tion system. 1. Introduction to methodology and encoding rules.Jour- nal of Chemical Information and Computer Sciences28, 1 (1988), 31–36. arXiv:https://doi.org/10.1021/ci00057a005 doi:10.1021/ci00057a005

    work page doi:10.1021/ci00057a005 1988

  47. [47]

    Eckart Zitzler and Lothar Thiele. 1998. Multiobjective Optimization Using Evolu- tionary Algorithms - A Comparative Case Study. InProceedings of the 5th Inter- national Conference on Parallel Problem Solving from Nature (PPSN V). Springer- Verlag, Berlin, Heidelberg, 292––304

    work page 1998

  48. [48]

    Zitzler, L

    E. Zitzler, L. Thiele, M. Laumanns, C.M. Fonseca, and V.G. da Fonseca. 2003. Performance assessment of multiobjective optimizers: an analysis and review. IEEE Transactions on Evolutionary Computation7, 2 (2003), 117–132. doi:10.1109/ TEVC.2003.810758 9 A Theoretical Foundations This appendix provides background on the quantum chemical and computational me...

    work page arXiv 2003