Pith. sign in

REVIEW 3 major objections 5 minor 56 references

$\texttt{codesign-mcdp}$: A Python Library for Monotone Co-Design Problems

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read A Python library implements the monotone co-design framework: design problems are monotone relations from functionality to resources, composed by series, parallel, and feedback, and solved to exact minimal-resource Pareto fronts by Kleene i

desk verdict Useful Python reimplementation of Censi's co-design calculus, but the genuinely new Section 15 layers rest on unproved theorems and a discretization that can violate their hypotheses; the core is solid, the novelty not yet. read the letter →

arxiv 2607.18415 v1 pith:YSRQA3ES submitted 2026-07-20 math.OC cs.MScs.SYeess.SY

Corentin Briat This is my paper · ORCID
classification math.OCcs.MScs.SYeess.SY MSC 90C2906A0647H10
keywords monotoneco-designantichaincalculusParetofrontKleenefixedpointcompositionoperatorssequentialPythonlibrarydesignautomation
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 manual presents codesign-mcdp, a from-scratch Python library that implements the monotone co-design framework: design problems are monotone relations from functionality posets to antichains of resources, composed by series, parallel, and feedback operators, and solved by a Kleene fixed-point iteration over the antichain lattice. The central claim is that the implementation is faithful—cyclic constraints (such as a battery that must lift its own mass) converge to the correct minimal-resource Pareto front—and that the library's temporal, sequential, vector-state, and online layers, which are the author's own work, work as documented. A sympathetic reader would care because this turns mutually dependent component specifications into a routine solve, rather than a bespoke fixed-point argument.

What carries the argument

The central object is a design problem: a monotone map h from a functionality poset F to the lattice of antichains of a resource poset R. The carrying mechanism is the antichain lattice ordered by domination, the three composition operators (series, parallel, feedback), and the Kleene iteration that closes feedback loops by ascending from the bottom antichain through the operator Phi_f(A) = Min(union over r in A of h(f, r_X) intersect up(r)) until a fixed point is reached. All six primitive problem types—algebraic, functional, catalog, constraint, ODE-derived, and uncertainty-bracketed—reduce to this same h: F -> A[R] interface.

What would settle it

Solve a small two-stage sequential co-design problem with a transition that is not jointly monotone (e.g., a state update with an interior optimum) on a fine grid, and compare the solver's value antichain to exhaustive enumeration of all policy choices; if any non-dominated total-resource point is missing, Q2 fails. For the core solver, re-running the drone example for several payloads p and checking that the converged battery mass equals 2p within numerical tolerance would confirm the Kleene implementation.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims that the antichain calculus, six primitive design-problem types, the three composition operators, the Kleene solver, and the two high-level builders constitute a complete, independent implementation of monotone co-design, so that composed problems with cycles are solved to the least fixed point, i.e., the minimal-resource Pareto front. The touchstones are a drone example where Kleene iteration converges to the closed-form fixed point m* = 2p, and an integer-optimization example whose published solution set is reproduced (with a flagged correction to that published set). The paper further claims that the temporal, sequential, vector-state, and online layers

Load-bearing premise

That the unproved results Q1–Q3—monotone value, front-equals-reachable-frontier, and exact factorization at resets—actually hold, since the sequential, vector-state, and dynamic layers rest on them; the manual supplies only sampled numerical checks, not proofs.

Editorial extensions

If this is right

  • Any monotone design problem can be composed and solved with a single call to solve, so cyclic resource–functionality couplings become a closed-box operation rather than a bespoke fixed-point derivation.
  • Feedback loops converge to the least fixed point (minimal resources), with infeasibility detected when the loop axis saturates to top.
  • Uncertainty can be reported as worst-case, mean, percentile, or CVaR in one solve, bracketed by user-supplied optimistic and pessimistic variants.
  • Online learning with certified confidence-polytope bounds eliminates provably suboptimal catalog candidates before their inner solve, reducing the number of full evaluations.
  • Multi-stage, state-dependent co-design is available through temporal (Viterbi) and sequential/vector-state (Bellman) planners, with exact schedules when the stated hypotheses hold.

Reading between the lines

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

  • If Q1–Q3 are eventually proved, the sequential layer gives a principled decomposition of finite-horizon design at state resets, analogous to regeneration-point decomposition in stochastic processes.
  • The dependency-free core makes it natural to embed co-design problems in larger optimization or learning pipelines, where the composition operators could serve as reusable modules.
  • A concrete test: combine the certified linear-parametric evaluator with the monotonicity evaluator on an affinely parameterized catalog with state-dependent feasibility; if the monotone part prunes the low-feature corner, evaluation counts should drop below either approach alone.
  • The library's non-verification of user-supplied monotonicity suggests an automatic monotonicity checker as a natural added safety layer for production use.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript is a reference manual for codesign-mcdp, a from-scratch Python implementation of Censi's monotone co-design framework. It claims to implement the antichain calculus, six primitive design-problem types, series/parallel/feedback composition, a Kleene fixed-point solver, an MCDPL-style declarative builder, and a modular System builder, together with uncertainty, online-learning, visualisation, and temporal/sequential/vector/online layers. Attribution is explicit: the core theory and the uncertainty and online-learning layers are credited to Censi and Alharbi et al., and the manual states plainly that the library 'contributes software, not theory.' The only claimed original content is Section 15, whose correctness is stated to depend on three unpublished results, Q1–Q3. Worked examples reproduce Censi's examples, correct one published output (the M(1) set in Example 2), and demonstrate the cyclic drone problem converging to the closed-form fixed point m*=2p.

Significance. If correct, the library is a genuinely useful reference implementation: it is independent, reproducible, well documented, ships 25 runnable examples and notebooks, and even provides a concrete check on a published result by flagging the M(1) typo. The core algorithmic claims are credible from the worked outputs and appear to be rendered faithfully. The original temporal, sequential, vector-state, and online layers would be a meaningful extension of monotone co-design, but their significance is conditional on Q1–Q3, which are not proved in the manuscript and are implemented over a discretisation that can violate their hypotheses. The manuscript's explicit attribution boundary is a notable strength, as is the availability of traceable solver outputs.

major comments (3)
  1. [§13.5.3, §15] The original contribution of the manuscript is Section 15, but its correctness rests on three results—monotone value (Q1), front-equals-reachable-frontier (Q2), and exact factorization at resets (Q3)—that are stated in §13.5.3 without proof and described in the Introduction as 'being written up separately and are not yet published.' The only guard is check_monotonicity, which samples at most max_violations=8 witnesses and is not a verification. Because these layers are the paper's own claimed contribution, it is not sufficient to defer the proofs to a separate write-up; please either include the proofs, or explicitly re-label the sequential/vector/dynamic layers as experimental and remove unqualified correctness claims such as 'makes three theory results operational.'
  2. [§13.5.2, §13.5.3] The shipped discretisation does not preserve the hypotheses of Q1–Q3. In solve_dynamic/solve_sequential, a transition that leaves the grid envelope is rejected before snapping, with the stated rationale that an over-spent resource is never silently rescued. For a consumable resource with the standard increasing_is_larger=True order, this makes the one-step value map non-monotone at the boundary: h(x_low)=∅ (the top of the antichain lattice) for states below the consumption threshold, while h(x_high) is finite for states above it, so h(x_low) is not ≤ h(x_high). Consequently the monotone-value guarantee (Q1) does not hold for the algorithm actually shipped, and Q2's 'front equals reachable frontier' is not established. The manual concedes that 'snapping is not order-preserving at bucket boundaries' but gives no quantitative bound. Please either modify the rejection/snapping to be order-pr
  3. [§5.1, §2.5] The central convergence guarantee of Theorem 2.13 applies only to monotone relations, but the library does not verify monotonicity: §5.1 states that 'Monotonicity is the caller's responsibility and is not verified.' A user who accidentally supplies a non-monotone AlgebraicDP or FunctionDP can silently obtain an output that is not the least fixed point, with no warning from the solver. This is a limitation rather than an internal inconsistency, but it should be prominent in the top-level API description and ideally accompanied by a cheap monotonicity diagnostic, since Section 17 already gives modelling guidelines where such a check would fit.
minor comments (5)
  1. [§14.2] The correction to Censi's M(1) is confirmed by direct calculation; a one-line derivation of why (1,0) and (0,1) are infeasible would strengthen the manual's pedagogical value.
  2. [§13.5.3] The phrase 'Spurious violations most often signal a too-coarse grid' is speculative without a concrete example; a small fabricated grid that shows a spurious violation would make the guidance actionable.
  3. [§13.6.1] The state_get error message example names the missing axis 'charge' while the snippet's state vector carries axes ['flag','fuel']; adjust the example so the message matches the surrounding vectors.
  4. [§12.1] Code listings use the Unicode symbol '≥' in place of Python's '>='; the text should note explicitly that the operator is sugar for the ASCII form so readers copying the snippets are not confused.
  5. [§1, §15] The Introduction mentions 'receding-horizon' layers, but Section 15's enumeration lists temporal, dynamic, sequential, vector, precompute-then-DP, and online feedback cases. Align the terminology so the advertised scope matches the documented API.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: theory is attributed to external authors, the implementation is checked against external benchmarks, and the unproved Q1-Q3 are disclosed as unpublished theory, not used as inputs disguised as predictions.

full rationale

The manual is structurally anti-circular. It explicitly assigns the theory of Section 2 to Censi ('The definitions, the theorems, and the fixed-point algorithm are all his; nothing in this section is original to this library') and describes the library as 'an independent Python implementation of that framework. It contributes software, not theory.' There are no fitted parameters whose values presuppose the claimed outputs: the Kleene solver iterates user-supplied relations, and the worked examples are checked against external sources (Censi's examples 1-5; Alharbi et al.'s example 25). The only original content, Section 15's temporal/sequential/vector layers, is expressly conditioned on Q1-Q3, which the Introduction states 'are being written up separately and are not yet published, so they carry no citation here.' That is an honest disclosure of an unproved premise, not a circular reduction: the manual does not claim Q1-Q3 follow from the library, and it does not invoke a prior theorem by the same author as a substitute for proof. The stated caveats - 'Monotonicity is the caller's responsibility and is not verified' and 'snapping is not order-preserving at bucket boundaries' - are assumptions and limitations rather than inputs engineered to force the predictions. Even if Q1-Q3 fail, the failure would be an unsupported correctness claim about the Section 15 layers, not a case of the derivation being equivalent to its inputs by construction. Accordingly, no specific circular step can be quoted, and the appropriate score is 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No free parameters are fitted anywhere: the paper makes no empirical claim calibrated to data. The example constants (specific energy 1.8e6 J/kg, heat-loss k=0.5, Lipschitz L values) are problem statements or documented user choices, not fitted values. The load-bearing axioms are: (1) Censi's fixed-point theorem (external, standard); (2) unverified monotonicity of user-supplied DPs; (3) the unproved Q1–Q3 results behind the Section 15 layers; (4) the modeling assumption that the shipped posets faithfully represent engineering quantities; (5) unexamined numerical convergence in ODE_DP. No invented scientific entities are introduced: the hidden '__modules__' bundled feedback axis in §12.4 is an implementation mechanism, not a scientific postulate, so it is not listed as an invented entity.

assumptions (5)
  • standard math Kleene iteration over the antichain lattice converges to the least fixed point of a monotone operator (Censi 2015, Prop. 4).
    Theorem 2.13, Section 2.5. External published result, parameter-free; the solver's termination conditions (fixed point, empty antichain, loop-axis top) inherit the proof.
  • domain assumption User-supplied design-problem relations are monotone; the library does not verify this.
    Section 5.1 (AlgebraicDP): "Monotonicity is the caller's responsibility and is not verified." The composition closure theorem and the fixed-point guarantee break if violated. Also acknowledged for FunctionDP ('the caller guarantees monotonicity').
  • ad hoc to paper The Q1–Q3 results hold: monotone value (Q1), front-equals-reachable-frontier (Q2), exact factorization at resets (Q3).
    Stated in Section 13.5.3 without proof; the Introduction says these results "are being written up separately and are not yet published, so they carry no citation here." They underpin solve_sequential, check_monotonicity, detect_resets, factorise_at_resets, and dp_over_catalog.
  • domain assumption The poset embeddings (Reals, Naturals, Ports, Discrete) faithfully represent the engineering quantities of interest.
    All examples use lumped-parameter physics (e.g., the drone relation m ≥ (c(g(m+m_pay))^2 + P_ex)T/α in §14.1). Modeling error from lumping is not quantified; the library computes within the posets, not over the real system.
  • domain assumption The numerical methods in ODE_DP (explicit Euler with default n_steps=200; Newton iteration) converge to the correct resource values.
    Section 5.5 and Example 5.4. No error analysis or convergence conditions are given; Example 5.4 asserts the steady-state root 'is recovered exactly by the Newton iteration.'

how reviews work

0 comments
Cite this review

Pith. "Pith review of $\texttt{codesign-mcdp}$: A Python Library for Monotone Co-Design Problems." pith.science (2026). https://pith.science/paper/YSRQA3ES

@misc{pith2026260718415,
  author       = {Pith},
  title        = {Pith review of: $\textttcodesign-mcdp$: A Python Library for Monotone Co-Design Problems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YSRQA3ES}},
  note         = {Machine review of arXiv:2607.18415}
}
abstract

$\texttt{codesign-mcdp}$ is a Python library for formulating and solving $\textit{Monotone Co-Design Problems}$ (MCDPs) in the framework of Censi (2015). A design problem is a relation between two posets, a functionality poset $F$ and a resource poset $R$; given a target functionality, the problem asks for the antichain of minimal resources needed to deliver it. Design problems compose under three operators (series, parallel, feedback), and the resulting class is closed under composition. The library implements the antichain calculus, six primitive design-problem types, the three composition operators, a Kleene fixed-point solver, and two high-level builders (an MCDPL-style declarative builder and a modular $\texttt{System}$ builder). Further layers add set-based and stochastic uncertainty, compositional online learning, and temporal, vector-state, and online co-design, alongside a suite of worked examples.

Figures

Figures reproduced from arXiv: 2607.18415 by the authors.

Figure 1
Figure 1. Kleene ascent for c = 4: each panel is one iterate Sk of the antichain in N × N, from the seed {(0, 0)} to the five-point fixed point M(4). Reading The panels make the antichain algebra tangible: points appear along a frontier x + y = const, the frontier advances outward, and then dominated points fall away under Min until only the incomparable minimal pairs remain. The figure is the exact geometric content of the S… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 1 canonical work pages

  1. [1]

    Modular Self-Configurable Robots—The State of the Art.Actuators, 12 (9):361, 2023

    Actuators survey. Modular Self-Configurable Robots—The State of the Art.Actuators, 12 (9):361, 2023. Floor-cleaning tiling robots hTrihex/hTetro and reconfiguration for coverage

    2023

  2. [2]

    Dahleh, and Gioele Zardini

    Meshal Alharbi, Munther A. Dahleh, and Gioele Zardini. Compositional Online Learning for Multi-Objective System Co-Design, 2026. URLhttps://arxiv.org/abs/2604.22624. arXiv:2604.22624

    arXiv 2026

  3. [3]

    X. An, Q. Jia, G. Chen, Y. Liu, and H. Liu. Morphological Design and Performance Analysis of a Self-reconfigurable Modular Robot for Multi-task Space Applications.Journal of Mechanical Engineering, 61(21):152–167, 2025

    2025

  4. [4]

    Optimization of Non-Equilibrium Self-Assembly Pro- tocols Using Markov State Models, 2022

    Authors as listed on the preprint. Optimization of Non-Equilibrium Self-Assembly Pro- tocols Using Markov State Models, 2022. URLhttps://arxiv.org/abs/2210.05749. arXiv:2210.05749; staged assembly-protocol optimisation

    arXiv 2022

  5. [5]

    Scalable Co-Design via Linear Design Problems: Compositional Theory and Algorithms, 2026

    Yubo Cai, Yujun Huang, Meshal Alharbi, and Gioele Zardini. Scalable Co-Design via Linear Design Problems: Compositional Theory and Algorithms, 2026. URLhttps://arxiv.org/ abs/2603.29083. arXiv:2603.29083

    arXiv 2026

  6. [6]

    Robot Co-design: Beyond the Monotone Case, 2019

    Luca Carlone and Carlo Pinciroli. Robot Co-design: Beyond the Monotone Case, 2019. URLhttps://arxiv.org/abs/1902.05880. arXiv:1902.05880; accepted at ICRA 2019. 119

    arXiv 2019

  7. [7]

    A Mathematical Theory of Co-Design, 2015

    Andrea Censi. A Mathematical Theory of Co-Design, 2015. URLhttps://arxiv.org/ abs/1512.08055. arXiv:1512.08055; last revised 2016. No journal version has appeared

    arXiv 2015

  8. [8]

    Monotone co-design problems; or, everything is the same

    Andrea Censi. Monotone co-design problems; or, everything is the same. In2016 American Control Conference (ACC). IEEE, 2016. doi: 10.1109/ACC.2016.7525085

    arXiv 2016

Show all 56 references
  1. [9]

    A Class of Co-Design Problems With Cyclic Constraints and Their Solu- tion.IEEE Robotics and Automation Letters, 2(1):96–103, 2017

    Andrea Censi. A Class of Co-Design Problems With Cyclic Constraints and Their Solu- tion.IEEE Robotics and Automation Letters, 2(1):96–103, 2017. doi: 10.1109/LRA.2016. 2535127

    2017 doi

  2. [10]

    Uncertainty in Monotone Codesign Problems.IEEE Robotics and Au- tomation Letters, 2(3):1556–1563, 2017

    Andrea Censi. Uncertainty in Monotone Codesign Problems.IEEE Robotics and Au- tomation Letters, 2(3):1556–1563, 2017. doi: 10.1109/LRA.2017.2674970. Preprint arXiv:1609.03103, where the title readsUncertainty in Monotone Co-Design Problems

    2017

  3. [11]

    MCDPL: the reference co-design software distribution, n.d

    Andrea Censi. MCDPL: the reference co-design software distribution, n.d. URLhttps:// co-design.science/software/. Accompanies the monotone co-design framework of Censi [7]; the distribution carries no release date

  4. [12]

    Applied Category Theory for Engi- neering, n.d

    Andrea Censi, Jonathan Lorand, and Gioele Zardini. Applied Category Theory for Engi- neering, n.d. URLhttps://applied-compositional-thinking.engineering/. Book in preparation, announced for Cambridge University Press; no publication date established

  5. [13]

    B. A. Davey and H. A. Priestley.Introduction to Lattices and Order. Cambridge University Press, 2nd edition, 2002

    2002

  6. [14]

    D. J. Docimo, Z. Kang, K. A. James, and A. G. Alleyne. Plant and Controller Optimization for Power and Energy Systems with Model Predictive Control.Journal of Dynamic Systems, Measurement, and Control, 2021

    2021

  7. [15]

    Composable Uncertainty in Symmetric Monoidal Categories for Design Problems

    Marius Furter, Yujun Huang, and Gioele Zardini. Composable Uncertainty in Symmetric Monoidal Categories for Design Problems. InProceedings of the Eighth International Con- ference on Applied Category Theory (ACT 2025), volume 442 ofEPTCS, pages 30–44, 2026. doi: 10.4204/EPTCS....

    2025

  8. [16]

    Gagnon, G

    M. Gagnon, G. Hiller, Y.-T. Luan, A. Kittredge, J. DeFelice, and D. Drapeau. High-end pH-controlled delivery of glucose (HIPDOG) effectively suppresses lactate accumulation in CHO fed-batch cultures.Biotechnology and Bioengineering, 108(6):1328–1337, 2011

    2011

  9. [17]

    Garcia-Sanz

    M. Garcia-Sanz. Control Co-Design: An engineering game changer.Advanced Control for Applications, 1(1):e18, 2019

    2019

  10. [18]

    Genta and L

    G. Genta and L. Morello.The Automotive Chassis, Vol. 1: Components Design. Springer,

  11. [19]

    D. R. Herber and J. T. Allison. Nested and Simultaneous Solution Strategies for General Combined Plant and Control Design Problems.Journal of Mechanical Design, 141(1): 011402, 2018

    2018

  12. [20]

    J. B. Heywood.Internal Combustion Engine Fundamentals. McGraw-Hill, 2nd edition,

  13. [21]

    Hofmann.Hybridfahrzeuge

    P. Hofmann.Hybridfahrzeuge. Springer, 2nd edition, 2014. Power-split topologies, sizing rules, efficiency models. 120

    2014

  14. [22]

    DistributionalUncertaintyandAdaptiveDecision-Making in System Co-design, 2026

    YujunHuangandGioeleZardini. DistributionalUncertaintyandAdaptiveDecision-Making in System Co-design, 2026. URLhttps://arxiv.org/abs/2603.14047. arXiv:2603.14047

    2026

  15. [23]

    On Composable and Parametric Uncertainty in Systems Co-Design, 2025

    Yujun Huang, Marius Furter, and Gioele Zardini. On Composable and Parametric Uncertainty in Systems Co-Design, 2025. URLhttps://arxiv.org/abs/2504.02766. arXiv:2504.02766; accepted as an invited session paper to IEEE CDC 2025

    2025 arXiv

  16. [24]

    M. C. Hübl and C. P. Goodrich. Accessing Semi-Addressable Self-Assembly with Efficient Structure Enumeration.Physical Review Letters, 134(5):058204, 2025

    2025

  17. [25]

    M. C. Hübl and C. P. Goodrich. Simultaneous optimization of assembly time and yield in programmable self-assembly.The Journal of Chemical Physics, 164(8):084904, 2026

    2026

  18. [26]

    Global EV Outlook

    International Energy Agency. Global EV Outlook. Technical report, International Energy Agency, 2023. Battery pack cost and pack-level energy-density trends, 2020–2030

    2023

  19. [27]

    S. F. Khattak, Z. Xing, B. Kenty, I. Koyrakh, and Z. J. Li. Feed development for a fed- batch CHO production process by semi-steady-state analysis.Biotechnology Progress, 26 (3):797–804, 2010

    2010

  20. [28]

    Lao and D

    C. Lao and D. Toth. Effects of ammonium and lactate on growth and metabolism of a recombinant CHO cell culture.Biotechnology Progress, 13:688–691, 1997

    1997

  21. [29]

    Larminie and J

    J. Larminie and J. Lowry.Electric Vehicle Technology Explained. Wiley, 2nd edition, 2012. Motor efficiency maps, inverter losses, charging-system architecture

    2012

  22. [30]

    Z. Liu, Q. Lu, J. Luo, and Z. Wang. Terrain-aware morphology searching algorithm for self-reconfigurable modular robot in dynamic environment.Applied Soft Computing, 186: 114182, 2026

    2026

  23. [31]

    CODEI: Resource-Efficient Task-Driven Co-Design of Perception and Decision Making for Mobile Robots Applied to Autonomous Vehicles.IEEE Transactions on Robotics, 41:2727–2748,

    Dejan Milojevic, Gioele Zardini, Miriam Elser, Andrea Censi, and Emilio Frazzoli. CODEI: Resource-Efficient Task-Driven Co-Design of Perception and Decision Making for Mobile Robots Applied to Autonomous Vehicles.IEEE Transactions on Robotics, 41:2727–2748,

  24. [32]

    Mock et al

    P. Mock et al. European Vehicle Market Statistics Pocketbook 2023/24. Technical report, International Council on Clean Transportation, 2023. Mass-versus-mission scatter data for calibration

    2023

  25. [33]

    Naunheimer et al.Automotive Transmissions

    H. Naunheimer et al.Automotive Transmissions. Springer, 2nd edition, 2011. Gearbox losses, weights, costs

    2011

  26. [34]

    Nemry et al

    F. Nemry et al. Environmental Improvement of Passenger Cars. Technical report, JRC Scientific and Technical Reports, 2008. CO2 emission factors, fuel densities, well-to-wheel

    2008

  27. [35]

    Marc-Philippe Neumann, Raphael Habermacher, Giona Fieni, Alberto Cerofolini, Gioele Zardini, and Christopher H. Onder. Hierarchical Co-Design for Multi-Race Strategy Opti- mization in Formula 1, 2025. ETH Zurich Research Collection working paper; to appear at the 29th IEEE Int...

    2025

  28. [36]

    W. W. Pulkrabek.Engineering Fundamentals of the Internal Combustion Engine. Pearson, 2nd edition, 2003. Thermal efficiency, heat rejection, fuel flow rates. 121

    2003

  29. [37]

    Reinhart, L

    D. Reinhart, L. Damjanovic, C. Kaisermayer, W. Sommeregger, A. Gili, B. Gasselhuber, et al. Bioprocessing of recombinant CHO-K1, CHO-DG44, and CHO-S: CHO expression hosts favor either mAb production or biomass synthesis.Biotechnology Journal, 14(3): e1700686, 2019

    2019

  30. [38]

    Quantale- Enriched Co-Design: Toward a Framework for Quantitative Heterogeneous System Design,

    Hans Riess, Yujun Huang, Matthew Klawonn, Gioele Zardini, and Matthew Hale. Quantale- Enriched Co-Design: Toward a Framework for Quantitative Heterogeneous System Design,

  31. [39]

    Bosch, 10th edition, 2018

    Robert Bosch GmbH.Bosch Automotive Handbook. Bosch, 10th edition, 2018. Engine specific power, exhaust aftertreatment, electrical loads

    2018

  32. [40]

    S. N. Sou, C. Sellick, K. Lee, et al. How does mild hypothermia affect monoclonal antibody glycosylation?Biotechnology and Bioengineering, 112(6):1165–1176, 2015

    2015

  33. [41]

    Sun et al

    J. Sun et al. Robust Control Co-Design with Receding-Horizon MPC, 2021. URLhttps: //arxiv.org/abs/2104.02025. arXiv:2104.02025; the online-feedback co-design pattern

    2021 arXiv

  34. [42]

    Trummer, K

    E. Trummer, K. Fauland, S. Seidinger, et al. Process parameter shifting: Part I. Effect of DOT, pH, and temperature on the performance of Epo-Fc expressing CHO cells cultivated in controlled batch bioreactors.Biotechnology and Bioengineering, 94(6):1033–1044, 2006

    2006

  35. [43]

    Environmental Protection Agency

    U.S. Environmental Protection Agency. Automotive Trends Report. Technical report, U.S. Environmental Protection Agency, 2024. Fleet-average fuel economy, weight-class data, drag-coefficient data

    2024

  36. [44]

    Various. Multi-objective protein design: representative works, 2017.Searching for the Pareto frontier in multi-objective protein design, Biophysical Reviews 9:311–324, 2017; Mo- sPro,Pareto-optimal sampling for multi-objective protein sequence design, 2025; ST-PARM and MAProt,...

    2017

  37. [45]

    Self-reconfiguring modular robots: overview, 2024

    Various. Self-reconfiguring modular robots: overview, 2024. SeeModular reconfigurable robots: Toward on-demand multifunctional applications, Science Robotics, 2024, and the surveyModular Self-Configurable Robots—The State of the Art, Actuators 12(9):361, 2023

    2024

  38. [46]

    S. K. Yoon, S. H. Kim, and G. M. Lee. Effect of low culture temperature on specific productivity and transcription level of anti-4-1BB antibody in recombinant Chinese hamster ovary (CHO) cells.Biotechnology Progress, 19:1383–1386, 2003

    2003

  39. [47]

    On the Co-Design of AV-Enabled Mobility Systems

    Gioele Zardini, Nicolas Lanzetti, Mauro Salazar, Andrea Censi, Emilio Frazzoli, and Marco Pavone. On the Co-Design of AV-Enabled Mobility Systems. In2020 IEEE 23rd Inter- national Conference on Intelligent Transportation Systems (ITSC), pages 1–8, 2020. doi: 10.1109/ITSC45102....

    2020

  40. [48]

    Towards a Co-Design Framework for Future Mobility Systems

    Gioele Zardini, Nicolas Lanzetti, Mauro Salazar, Andrea Censi, Emilio Frazzoli, and Marco Pavone. Towards a Co-Design Framework for Future Mobility Systems. In99th An- nual Meeting of the Transportation Research Board, 2020. doi: 10.3929/ethz-b-000373705. Preprint arXiv:1910.07714

    2020 arXiv

  41. [49]

    Co-Design of Autonomous Sys- tems: From Hardware Selection to Control Synthesis

    Gioele Zardini, Andrea Censi, and Emilio Frazzoli. Co-Design of Autonomous Sys- tems: From Hardware Selection to Control Synthesis. In2021 European Control Con- ference (ECC), pages 682–689, 2021. doi: 10.23919/ECC54610.2021.9654960. Preprint arXiv:2011.10758. 122

    2021

  42. [50]

    Co-design of Embodied Intelligence: A Structured Approach

    Gioele Zardini, Dejan Milojevic, Andrea Censi, and Emilio Frazzoli. Co-design of Embodied Intelligence: A Structured Approach. In2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 7536–7543, 2021. doi: 10.1109/IROS51168. 2021.9636513. Prepri...

    2021

  43. [51]

    Task-driven Modular Co-design of Vehicle Control Systems

    Gioele Zardini, Zelio Suter, Andrea Censi, and Emilio Frazzoli. Task-driven Modular Co-design of Vehicle Control Systems. In2022 IEEE 61st Conference on Decision and Control (CDC), pages 2196–2203, 2022. doi: 10.1109/CDC51059.2022.9993107. Preprint arXiv:2203.16640

    2022

  44. [52]

    Co- Design to Enable User-Friendly Tools to Assess the Impact of Future Mobility Solutions

    Gioele Zardini, Nicolas Lanzetti, Andrea Censi, Emilio Frazzoli, and Marco Pavone. Co- Design to Enable User-Friendly Tools to Assess the Impact of Future Mobility Solutions. IEEE Transactions on Network Science and Engineering, 10(2):827–844, 2023. doi: 10. 1109/TNSE.2022.322...

    2023

  45. [2009]

    Suspension stiffness, damper characteristics, brake sizing, tire load ratings

  46. [2018]

    Combustion thermodynamics, BSFC ranges, durability

  47. [2025]

    Preprint arXiv:2503.10296

    doi: 10.1109/TRO.2025.3552347. Preprint arXiv:2503.10296

    2025

  48. [2026]

    arXiv:2603.29921

    URLhttps://arxiv.org/abs/2603.29921. arXiv:2603.29921

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.