Pith. sign in

REVIEW 3 major objections 4 minor 57 references

Direct Pseudospectral Optimal Control by Orthogonal Polynomial Integral Collocation

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A new direct pseudospectral transcription, orthogonal polynomial integral collocation, represents only the highest-order derivative of each coordinate and integrates it to recover all lower state levels, reducing the NLP decision-variable…

desk verdict The modal integral-collocation idea is sound, but the Chebyshev quadrature weights are misprinted, which blocks reproduction until fixed. read the letter →

arxiv 2505.19454 v2 pith:DTU54FAK submitted 2025-05-26 math.OC cs.SYeess.SY

Thomas L. Ahrens , Ian M. Down , Manoranjan Majji This is my paper · ORCID
classification math.OCcs.SYeess.SY MSC 49M3765L6065D32
keywords directpseudospectralmethodsorthogonalpolynomialintegralcollocationChebyshevpolynomialsLegendretrajectoryoptimizationnonlinearprogrammingtranscriptionrocketlandingflipmaneuver
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 introduces a direct transcription method, orthogonal polynomial integral collocation (OPIC), for converting a continuous-time optimal control problem into a nonlinear program. Rather than representing each state level on a grid and using differentiation matrices to enforce the dynamics, OPIC expands only the highest-order derivative of each coordinate in an orthogonal polynomial series and obtains every lower derivative level by repeated exact integration. The result is a modal formulation in which one set of polynomial coefficients carries an entire degree of freedom, so the decision-variable vector of the NLP is much shorter than in standard nodal pseudospectral methods. The paper verifies the method on the minimum-fuel and Breakwell benchmark problems with analytical solutions, reproduces published planar circular orbit-raising transfer times, and solves a constrained Starship-style landing flip maneuver. A consistent finding across examples is that the choice of polynomial basis and associated grid materially affects runtime, accuracy, and convergence.

What carries the argument

The mechanism is the integration-weight matrix chain. For each polynomial family, the indefinite integrals of the basis functions close onto adjacent basis functions, $\int \phi_i = b_i^+ \phi_{i+1} + b_i^- \phi_{i-1}$, so integrating the modal series $\sum \phi_i \alpha_i$ is exactly a sparse matrix product $B_j$ with entries only on the super- and sub-diagonals. Repeatedly applying $B_j$ and subtracting the lower-bound evaluation vectors $v_j$ reconstructs every derivative of the state from the single vector $\alpha$, while initial conditions enter through the polynomial terms $p_{m-j}(\tau)$. This single-set-per-coordinate property is what replaces the state-ladder variables of nodal pseudospectral methods and drives the reduction in NLP dimension.

What would settle it

Take a bang-bang problem with a known switching time that does not coincide with any collocation node, compute the maximum deviation between the OPIC polynomial and the true highest-order derivative as $n$ increases, and also evaluate the state constraint at off-node points. If the error fails to decay spectrally or the constraint is violated between nodes, as the odd-order Breakwell solution already suggests can happen, the single-global-polynomial assumption is falsified.

Watch

Extended reading notes

Core claim

The central claim is that a single $n$th-order orthogonal polynomial series can represent the highest-order derivative $\ddot{y}$ (or $y^{(q)}$) of each coordinate, and that successive integrations of that series, carried out exactly through sparse integration-weight matrices $B_j$ built from the polynomial recurrences, yield all lower derivative levels with no additional coefficients. Collocation enforces the dynamics at the nodes, boundary conditions are imposed on the integrated state expressions, and controls stay nodal so that discontinuous or bang-bang control can still be represented. With this construction, the state part of the NLP decision vector contains only the $n+1$ modal coefficients $\alpha$ per coordinate instead of a full ladder of nodal state values, which is the paper's main efficiency claim. The numerical experiments show that all tested bases, Chebyshev of the first and second kind and Legendre on their associated grids, reproduce the analytical benchmarks, match the orbit-raising transfer time of about $t_f \simeq 3.32$ TU reported by earlier pseudospectral studies, and yield similar landing flip solutions, with second-kind Chebyshev often fastest and Legendre grids most reliable.

Load-bearing premise

The load-bearing premise is that one global polynomial, fixed in advance and spanning the whole time interval, can faithfully capture the highest-order derivative of the trajectory, even when that derivative contains sharp switches or corners; the method does not adapt by splitting the domain.

Editorial extensions

If this is right

  • For a second-order state, the state part of the NLP decision vector drops from roughly $2(n+1)$ nodal values to $n+1$ modal coefficients, with the gap growing for higher-order dynamics.
  • Because integration is a smoothing operation, derivative and state histories produced from the integrated series should be less error-amplifying than those obtained by differentiating nodal state values.
  • The method inherits spectral accuracy on smooth problems, so on trajectories whose highest derivative is smooth, increasing $n$ should drive the transcribed solution toward the continuous optimum quickly.
  • Control stays nodal, so the modal state representation does not prevent representing piecewise or bang-bang control histories.
  • Polynomial and grid choice is not neutral: on the tested problems, second-kind Chebyshev gave the lowest average runtime for the minimum-fuel case while Legendre LG and LGL grids gave the highest convergence rates on orbit raising.

Reading between the lines

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

  • If domain segmentation were added, the modal coefficient vector per segment would become the natural adaptation target, allowing piecewise error near switching times to be reduced locally; nothing in the integral formulation prevents this extension.
  • The even-odd Breakwell discrepancy suggests a practical rule for users: when a state constraint may be active, prefer grids that include the constraint boundary or add off-node constraint checks, because node-only enforcement can leave the constraint violated between nodes.
  • The parameter-space advantage should be largest for high-order dynamics; a decisive head-to-head test would compare DOPIC with Gauss, Radau, and Legendre nodal methods on a fourth-order system at equal node counts, reporting decision-variable counts, accuracy, and solver time.
  • The paper leaves costate recovery open, but because the state is a global modal expansion, the NLP Lagrange multipliers could plausibly be mapped back to the indirect problem via a covector mapping theorem, turning DOPIC into a warm start for shooting methods.
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

3 major / 4 minor

Summary. The paper proposes Direct Orthogonal Polynomial Integral Collocation (DOPIC), a modal direct pseudospectral transcription of optimal control problems. The highest derivative of each state is represented by a single Chebyshev or Legendre polynomial series, and all lower derivative levels are obtained by repeated exact integration of that series, so that only one vector of modal coefficients is needed per coordinate. The Bolza cost is evaluated by quadrature, controls are represented nodally, and the resulting NLP is solved with fmincon. The method is demonstrated on a second-order minimum-fuel problem, the Breakwell bounded-state problem, planar circular orbit raising (minimum-time and maximum-radius variants), and a simplified Starship landing flip maneuver. The authors conclude that the single-coefficient-set representation reduces the parameter space and that the examples validate the method.

Significance. If the formulas as printed were correct, DOPIC would be a useful and reasonably general modal alternative to the established nodal pseudospectral methods: it applies to Chebyshev and Legendre bases with their standard grids, avoids rewriting dynamics in first-order form, and the benchmark results against analytical values and published orbit-raising transfer times are encouraging. The paper also documents several nontrivial limitations honestly, including the even/odd Breakwell artifact and the restriction to global collocation. However, the printed quadrature weights for the Chebyshev and CP2K node sets are not valid integration rules, and because Eq. (68) evaluates the cost with those weights, the reported performance indices cannot be reproduced from the manuscript as written. This defect is load-bearing for the central validation and must be fixed. The absence of any convergence theorem or error analysis also means the method's validity currently rests entirely on numerical evidence.

major comments (3)
  1. [II.B (Eqs. 31–33) and III.C (Eq. 68)] The quadrature weight formulas for the Chebyshev-Gauss, Chebyshev-Gauss-Lobatto, and CP2K nodes are internally inconsistent and do not define valid quadrature rules. For n=2, Eqs. (32a)–(32b) give w0=w2=1/4 and w1=7/3, which sum to 17/6 instead of 2; the rule therefore does not integrate f(tau)=1 exactly over [-1,1]. The correct Clenshaw-Curtis weights for these three nodes are w0=w2=1/3 and w1=4/3. The same defect appears in Eq. (31): for n=1 the two CG nodes each receive weight 2, again summing to 4 instead of 2, and Eq. (33) similarly gives weight 2 at each of the two CP2K nodes for n=1. Since Eq. (68) uses these weights to approximate the Bolza cost, an independent implementation following the printed transcription will compute incorrect performance indices and cannot reproduce the reported values such as J=14.142 in Fig. 2 or J=2.240 in Fig. 4. Please correct the weight formulas (or clearly state which different weights the actual code used) and re-run the affected examples.
  2. [IV.B (Eq. 78)] The Breakwell results for odd approximation orders are presented as valid solutions even though the continuous state constraint x(t) <= l is enforced only at the collocation nodes. As the paper itself notes, odd orders place no node at the constraint peak, so the constraint is inactive at every node and the reported cost underestimates the analytical value; in fact the polynomial state history violates the continuous constraint between nodes. This is not merely a cosmetic artifact: it means the odd-order performance indices in Fig. 4 correspond to infeasible trajectories. Please either enforce the constraint on a denser set of points, add a quantitative statement of maximum inter-node constraint violation, or explicitly exclude odd-order solutions from the accuracy claims.
  3. [III.C and V] The paper offers no convergence theorem or error bound for the DOPIC transcription, and it explicitly restricts the presentation to global collocation (III.C). For a new transcription method, the central claim that DOPIC is a valid and accurate pseudospectral method would be materially strengthened by at least a convergence statement for smooth problems and a discussion of how the polynomial order relates to the NLP solution error. In the current manuscript, after the quadrature issue is fixed, the only support for the central claim is the numerical evidence; the Starship problem is additionally verified only by internal consistency across node sets, since no external solution is provided. Please add a formal or at least quantitative convergence discussion, and state clearly which of the numerical claims are validated externally and which are not.
minor comments (4)
  1. [II.A (Eq. 14)] The recurrence for Chebyshev polynomials of the second kind is misprinted: U_{n+1}(tau) = 2 tau U_n(tau) - U_n(tau) should read U_{n+1}(tau) = 2 tau U_n(tau) - U_{n-1}(tau).
  2. [II.A] There is a spelling error in 'Chebyshev-Guass-Lobatto' (should be 'Gauss'), and Section IV contains the duplicated phrase 'All computations are are run'.
  3. [Fig. 2] The caption of Fig. 2 refers to the 'second order minimum control problem'; the problem is the second-order minimum-fuel problem. This should be made consistent.
  4. [III.A] The claim that 'no additional coefficients are needed for each integration' is slightly misleading because the integration matrices B_j grow in dimension at each integration level; the statement would be clearer as 'no additional independent coefficient vectors are introduced'.

Circularity Check

0 steps flagged · score 1.0 of 10

The DOPIC derivation is self-contained: its transcription follows from standard orthogonal-polynomial integral identities and is validated against external analytical and published solutions; the self-citations establish lineage but are not load-bearing.

full rationale

The paper does not exhibit a circular derivation chain. The core transcription in Section III reconstructs the integral-collocation matrices directly from standard integral identities for Chebyshev and Legendre polynomials (Eqs. 8-10, 19-21, 25-27, and Table 1) rather than importing the result from the cited ICC work. The validations in Section IV compare against external analytical solutions (minimum-fuel and Breakwell problems) and against other published pseudospectral solutions (orbit raising), so no fitted parameter is renamed as a prediction. The self-citations [27,32] are used for context and lineage, not as the proof of the central claim; the uniqueness-related or ansatz-smuggling patterns are absent. The only notable defect in the printed text is an apparent inconsistency in the CGL quadrature weights in Eqs. (32a)-(32b), which is a reproducibility/correctness issue, not a circularity issue. Overall, the central claim that a single polynomial-coefficient set can represent all derivative levels is a mathematical property of the integral formulation, and the benchmarks are independent of the formulation's own outputs.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

No new physical entities or fitted constants are introduced. The main user-chosen quantity is polynomial order n. Physical parameters in the orbit and Starship problems are borrowed from the problem definitions and public sources. The principal load-bearing assumptions are global polynomial smoothness of the highest derivative and node-only enforcement of constraints.

free parameters (1)
  • Polynomial approximation order n = 15-50 for the benchmark problems, 40 for orbit raising, 60 for the Starship flip
    Chosen by hand per problem. It controls accuracy and runtime, as shown in Figs. 2 and 4, but it is a standard discretization parameter rather than a constant fitted to the target solution.
assumptions (4)
  • standard math Orthogonal polynomial integral recurrences in Eqs. (8)-(10), (19)-(21), and (25)-(27) hold and integration constants cancel through lower-bound subtraction.
    This is the basis for the B matrices in Eqs. (42)-(55) and for the entire collocation formulation in Section III.A.
  • domain assumption The highest-order derivative of each coordinate can be represented accurately by one global nth-order polynomial over the entire time domain.
    Section III.C uses global collocation with one polynomial basis and explicitly leaves domain segmentation out of scope. The paper's own results show errors at bang-bang switching times and CP2K gimbal oscillations, confirming the assumption is load-bearing.
  • standard math Gaussian and Clenshaw-Curtis type quadrature weights in Eqs. (29)-(33) correctly discretize the continuous cost functional.
    Used in Eq. (68) to evaluate the integral cost. This is a standard quadrature result, but the paper does not quantify quadrature error for non-polynomial integrands.
  • domain assumption fmincon with finite-differenced Jacobians returns local solutions that accurately approximate the true local optima of the transcribed NLPs.
    Section IV states all problems are solved with fmincon SQP or interior-point methods. The 63-97 percent convergence rates in Tables 2 and 3 show the solver and initialization are material to the reported results.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Direct Pseudospectral Optimal Control by Orthogonal Polynomial Integral Collocation." pith.science (2026). https://pith.science/paper/DTU54FAK

@misc{pith2026250519454,
  author       = {Pith},
  title        = {Pith review of: Direct Pseudospectral Optimal Control by Orthogonal Polynomial Integral Collocation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DTU54FAK}},
  note         = {Machine review of arXiv:2505.19454}
}
read the original abstract

This paper details a methodology to transcribe an optimal control problem into a nonlinear program for generation of the trajectories that optimize a given functional by approximating only the highest order derivatives of a given system's dynamics. The underlying method uses orthogonal polynomial integral collocation by which successive integrals are taken to approximate all lower order states. Hence, one set of polynomial coefficients can represent an entire coordinate's degree of freedom. Specifically, Chebyshev polynomials of the first and second kind and Legendre polynomials are used over their associated common interpolating grids derived from the bases' roots and extrema. Simple example problems compare different polynomial bases' performance to analytical solutions. The planar circular orbit raising problem is used to verify the method with solutions obtained by other pseudospectral methods in literature. Finally, a rocket landing flip maneuver problem is solved to demonstrate the ability to solve complex problems with multiple states and control variables with constraints. Simulations establish this method's performance, and reveal that the polynomial/node choice for a given problem notably affects the performance.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 40 canonical work pages

  1. [1]

    E.,Applied optimal control: optimization, estimation and control, Routledge, 2018

    Bryson, A. E.,Applied optimal control: optimization, estimation and control, Routledge, 2018. https://doi.org/10.1201/ 9781315137667

    work page 2018

  2. [2]

    A.,Spacecraft trajectory optimization, Vol

    Conway, B. A.,Spacecraft trajectory optimization, Vol. 29, Cambridge University Press, 2010

    work page 2010

  3. [3]

    Direct trajectory optimization by a Chebyshev pseudospectral method,

    Fahroo, F., and Ross, I. M., “Direct trajectory optimization by a Chebyshev pseudospectral method,”Journal of Guidance, Control, and Dynamics, Vol. 25, No. 1, 2002, pp. 160–166. https://doi.org/10.2514/2.4862

    doi:10.2514/2.4862 2002

  4. [4]

    Connections between the covector mapping theorem and convergence of pseudospectral methods for optimal control,

    Gong, Q., Ross, I. M., Kang, W., and Fahroo, F., “Connections between the covector mapping theorem and convergence of pseudospectral methods for optimal control,”Computational Optimization and Applications, Vol. 41, 2008, pp. 307–335. https://doi.org/10.1007/s10589-007-9102-4

    doi:10.1007/s10589-007-9102-4 2008

  5. [5]

    Costate computation by a Chebyshev pseudospectral method,

    Gong, Q., Ross, I. M., and Fahroo, F., “Costate computation by a Chebyshev pseudospectral method,”Journal of Guidance, Control, and Dynamics, Vol. 33, No. 2, 2010, pp. 623–628. https://doi.org/10.2514/1.45154

    doi:10.2514/1.45154 2010

  6. [6]

    Direct–Indirect Hybrid Strategy for Optimal Powered Descent and Landing,

    Spada, F., Sagliano, M., and Topputo, F., “Direct–Indirect Hybrid Strategy for Optimal Powered Descent and Landing,”Journal of Spacecraft and Rockets, Vol. 60, No. 6, 2023, pp. 1787–1804. https://doi.org/10.2514/1.A35650

    work page doi:10.2514/1.a35650 2023

  7. [7]

    Survey of numerical methods for trajectory optimization,

    Betts, J. T., “Survey of numerical methods for trajectory optimization,”Journal of guidance, control, and dynamics, Vol. 21, No. 2, 1998, pp. 193–207. https://doi.org/10.2514/2.4231

    doi:10.2514/2.4231 1998

  8. [8]

    Survey of direct transcription for low-thrust space trajectory optimization with applications,

    Topputo, F., Zhang, C., et al., “Survey of direct transcription for low-thrust space trajectory optimization with applications,” Abstract and Applied Analysis, Vol. 2014, Hindawi, 2014. https://doi.org/10.1155/2014/851720

    doi:10.1155/2014/851720 2014

Show all 57 references
  1. [9]

    GPOPS-II: A MATLAB software for solving multiple-phase optimal control problems using hp-adaptive Gaussian quadrature collocation methods and sparse nonlinear programming,

    Patterson, M. A., and Rao, A. V., “GPOPS-II: A MATLAB software for solving multiple-phase optimal control problems using hp-adaptive Gaussian quadrature collocation methods and sparse nonlinear programming,”ACM Transactions on Mathematical Software (TOMS), Vol. 41, No. 1, 2014...

    2014 doi

  2. [10]

    Enhancements to the DIDO optimal control toolbox,

    Ross, I. M., “Enhancements to the DIDO optimal control toolbox,”arXiv preprint arXiv:2004.13112, 2020

    2004 arXiv

  3. [11]

    SPARTAN: A Novel Pseudospectral Algorithm for Entry, Descent, andLandingAnalysis,

    Sagliano, M., Theil, S., D’Onofrio, V., and Bergsma, M., “SPARTAN: A Novel Pseudospectral Algorithm for Entry, Descent, andLandingAnalysis,”AdvancesinAerospaceGuidance,NavigationandControl,editedbyB.Dołęga,R.Głębocki,D.Kordos, and M. Żugaj, Springer International Publishing, C...

    2018 doi

  4. [12]

    Solving complex optimal control problems at no cost with PSOPT,

    Becerra, V. M., “Solving complex optimal control problems at no cost with PSOPT,”2010 IEEE International Symposium on Computer-Aided Control System Design, 2010, pp. 1391–1396. https://doi.org/10.1109/CACSD.2010.5612676

    2010

  5. [13]

    B.,Chebyshev Polynomials in Numerical Analysis, Oxford University Press, 1968

    Fox, L., and Barker, I. B.,Chebyshev Polynomials in Numerical Analysis, Oxford University Press, 1968

    1968

  6. [14]

    A Chebyshev technique for solving nonlinear optimal control problems,

    Vlassenbroeck, J., and Van Dooren, R., “A Chebyshev technique for solving nonlinear optimal control problems,”IEEE transactions on automatic control, Vol. 33, No. 4, 1988, pp. 333–340. https://doi.org/10.1109/9.192187. 31

    1988 doi

  7. [15]

    Zero-propellant maneuver guidance,

    Bedrossian, N. S., Bhatt, S., Kang, W., and Ross, I. M., “Zero-propellant maneuver guidance,”IEEE Control Systems Magazine, Vol. 29, No. 5, 2009, pp. 53–73. https://doi.org/10.1109/MCS.2009.934089

    2009

  8. [16]

    A review of pseudospectral optimal control: From theory to flight,

    Ross, I. M., and Karpenko, M., “A review of pseudospectral optimal control: From theory to flight,”Annual Reviews in Control, Vol. 36, No. 2, 2012, pp. 182–197. https://doi.org/10.1016/j.arcontrol.2012.09.002

    2012 doi

  9. [17]

    ThepseudospectralLegendremethodfordiscretizingoptimalcontrolproblems,

    Elnagar,G.,Kazemi,M.A.,andRazzaghi,M.,“ThepseudospectralLegendremethodfordiscretizingoptimalcontrolproblems,” IEEE Transactions on Automatic Control, Vol. 40, No. 10, 1995, pp. 1793–1796. https://doi.org/10.1109/9.467672

    1995 doi

  10. [18]

    Legendre pseudospectral approximations of optimal control problems,

    Ross, I. M., and Fahroo, F., “Legendre pseudospectral approximations of optimal control problems,”New Trends in Nonlinear Dynamics and Control, and their Applications, Springer, 2004, pp. 327–342. https://doi.org/10.1007/978-3-540-45056-6_21

    2004 doi

  11. [19]

    Direct trajectory optimization and costate estimation via an orthogonal collocation method,

    Benson, D. A., Huntington, G. T., Thorvaldsen, T. P., and Rao, A. V., “Direct trajectory optimization and costate estimation via an orthogonal collocation method,”Journal of Guidance, Control, and Dynamics, Vol. 29, No. 6, 2006, pp. 1435–1440. https://doi.org/10.2514/1.20478

    2006 doi

  12. [20]

    A Gauss pseudospectral transcription for optimal control,

    Benson, D., “A Gauss pseudospectral transcription for optimal control,” Ph.D. thesis, Massachusetts Institute of Technology, 2005

    2005

  13. [21]

    Direct trajectory optimizationandcostateestimationoffinite-horizonandinfinite-horizonoptimalcontrolproblemsusingaRadaupseudospectral method,

    Garg, D., Patterson, M. A., Francolin, C., Darby, C. L., Huntington, G. T., Hager, W. W., and Rao, A. V., “Direct trajectory optimizationandcostateestimationoffinite-horizonandinfinite-horizonoptimalcontrolproblemsusingaRadaupseudospectral method,”Computational Optimization an...

    2011

  14. [22]

    An overview of three pseudospectral methods for the numerical solution of optimal control problems,

    Garg, D., Patterson, M., Hager, W., Rao, A., Benson, D. R., and Huntington, G. T., “An overview of three pseudospectral methods for the numerical solution of optimal control problems,” 2017

    2017

  15. [23]

    A Chebyshev approximation for solving optimal control problems,

    El-Gindy, T., El-Hawary, H., Salim, M., and El-Kady, M., “A Chebyshev approximation for solving optimal control problems,” Computers & Mathematics with Applications, Vol. 29, No. 6, 1995, pp. 35–45. https://doi.org/10.1016/0898-1221(95)00005-J

    1995 doi

  16. [24]

    A spectral method based on the second kind Chebyshev polynomials for solving a class of fractional optimal control problems,

    Nemati, S., “A spectral method based on the second kind Chebyshev polynomials for solving a class of fractional optimal control problems,”Sahand Communications in Mathematical Analysis, Vol. 4, No. 1, 2016, pp. 15–27

    2016

  17. [25]

    Solving boundary value problems, integral, and integro-differential equations using Gegenbauer integration matrices,

    Elgindy, K. T., and Smith-Miles, K. A., “Solving boundary value problems, integral, and integro-differential equations using Gegenbauer integration matrices,”Journal of Computational and Applied Mathematics, Vol. 237, No. 1, 2013, pp. 307–325. https://doi.org/10.1016/j.cam.2012.05.024

    2013 doi

  18. [26]

    Fast, accurate, and small-scale direct trajectory optimization using a Gegenbauer transcription method,

    Elgindy, K. T., and Smith-Miles, K. A., “Fast, accurate, and small-scale direct trajectory optimization using a Gegenbauer transcription method,”Journal of Computational and Applied Mathematics, Vol. 251, 2013, pp. 93–116. https://doi.org/10. 1016/j.cam.2013.03.032

    2013

  19. [27]

    Adaptive Collocation Methods Using Chebyshev Integration,

    Peck, C., “Adaptive Collocation Methods Using Chebyshev Integration,” Ph.D. thesis, Texas A&M University, 2023

    2023

  20. [28]

    Comparison of global and local collocation methods for optimal control,

    Huntington, G. T., and Rao, A. V., “Comparison of global and local collocation methods for optimal control,”Journal of guidance, control, and dynamics, Vol. 31, No. 2, 2008, pp. 432–436. https://doi.org/10.2514/1.30915. 32

    2008 doi

  21. [29]

    An hp-adaptive pseudospectral method for solving optimal control problems,

    Darby, C. L., Hager, W. W., and Rao, A. V., “An hp-adaptive pseudospectral method for solving optimal control problems,” Optimal Control Applications and Methods, Vol. 32, No. 4, 2011, pp. 476–502. https://doi.org/10.1002/oca.957

    2011 doi

  22. [30]

    P.,Chebyshev and Fourier spectral methods, Courier Corporation, 2001

    Boyd, J. P.,Chebyshev and Fourier spectral methods, Courier Corporation, 2001

    2001

  23. [31]

    https://doi.org/10.1201/9781420036114

    Mason, J., and Handscomb, D.,Chebyshev Polynomials, Chapman & Hall/CRC, 2003. https://doi.org/10.1201/9781420036114

    2003 doi

  24. [32]

    Direct Pseudospectral Optimal Control using Integral Chebyshev Collocation,

    Down, I. M., Peck, C. H., and Majji, M., “Direct Pseudospectral Optimal Control using Integral Chebyshev Collocation,”AIAA SCITECH 2024 Forum, 2024. https://doi.org/10.2514/6.2024-1461

    2024 doi

  25. [33]

    Modified Chebyshev-Picard Iteration Methods for Solution of Initial Value and Boundary Value Problems,

    Bai, X., “Modified Chebyshev-Picard Iteration Methods for Solution of Initial Value and Boundary Value Problems,” Ph.D. thesis, Texas A&M University, 2010

    2010

  26. [34]

    P., Brychkov, Y

    Prudnikov, A. P., Brychkov, Y. A., and Marichev, O. I.,Integrals and Series, Volume 2: Special Functions, Gordon and Breach Science Publishers, 1986

    1986

  27. [35]

    Inequalities for the Zeros of Legendre Polynomials and Related Functions,

    Szegö, G., “Inequalities for the Zeros of Legendre Polynomials and Related Functions,”Transactions of the American Mathematical Society, Vol. 39, No. 1, 1936

    1936

  28. [36]

    IQPACK: FORTRAN Subroutines for the Weights of Interpolary Quadratures,

    Elhay, S., and Kautsky, J., “IQPACK: FORTRAN Subroutines for the Weights of Interpolary Quadratures,”ACM Transactions on Mathematical Software, Vol. 13, No. 4, 1987, pp. 399–415

    1987

  29. [37]

    Fast and Accurate Computation of Gauss-Legendre and Gauss-Jacobi Quadrature Nodes and Weights,

    Hale, N., and Townsend, A., “Fast and Accurate Computation of Gauss-Legendre and Gauss-Jacobi Quadrature Nodes and Weights,”SIAM Journal on Scientific Computing, Vol. 35, No. 2, 2013, pp. 652–674. https://doi.org/10.1137/120889873

    2013 doi

  30. [38]

    Iteration-Free Computation of Gauss-Legendre Quadrature Nodes and Weights,

    Bogaert, I., “Iteration-Free Computation of Gauss-Legendre Quadrature Nodes and Weights,”SIAM Journal on Scientific Computing, Vol. 36, No. 3, 2014, pp. 1008–1026. https://doi.org/10.1137/140954969

    2014 doi

  31. [39]

    Methodus nova integralium valores per approximationem inveniende,

    Gauss, C. F., “Methodus nova integralium valores per approximationem inveniende,”Comment Soc. Reg. Scient. Gotting, 1814, pp. 39–76

  32. [40]

    Calculation of Gauss Quadrature Rules,

    Golub, G. H., and Welsch, J. H., “Calculation of Gauss Quadrature Rules,”Mathematics of Computation, Vol. 23, No. 106, 1969, pp. 221–230

    1969

  33. [41]

    A.,Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 10th ed., National Bureau of Standards, Applied Mathematics Series, 1972

    Abramowitz, M., and Stegun, I. A.,Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 10th ed., National Bureau of Standards, Applied Mathematics Series, 1972

    1972

  34. [42]

    Y., Quarteroni, A., and Jr., T

    Canuto, C., Hussaini, M. Y., Quarteroni, A., and Jr., T. A. Z.,Spectral Methods in Fluid Dynamics, 1st ed., Scientific Computation, Springer, 1988

    1988

  35. [43]

    Rabinowitz, P., and Davis, P.,Methods of Numerical Integration, 2nd ed., Dover, 2007

    2007

  36. [44]

    A method for numerical integration on an automatic computer,

    Clenshaw, C. W., and Curtis, A. R., “A method for numerical integration on an automatic computer,”Numerische Mathematik, Vol. 2, 1960, pp. 197–205. 33

    1960

  37. [45]

    On the infinite sequences arising in theories of harmonic analysis, of interpolation, and of mechanical quadrature,

    Fejér, L., “On the infinite sequences arising in theories of harmonic analysis, of interpolation, and of mechanical quadrature,” Bulletin of the American Mathematical Society, Vol. 39, 1933, pp. 521–534

    1933

  38. [46]

    Generation of higher order pseudospectral integration matrices,

    Elgindy, K. T., “Generation of higher order pseudospectral integration matrices,”Applied Mathematics and Computation, Vol. 209, No. 2, 2009, pp. 153–161. https://doi.org/10.1016/j.amc.2008.08.056

    2009 doi

  39. [47]

    E., and Ho, Y

    Bryson, A. E., and Ho, Y. C.,Applied Optimal Control: Optimization, Estimation, and Control, Taylor & Francis Group, 1975

    1975

  40. [48]

    Pseudospectral Chebyshev optimal control of constrained nonlinear dynamical systems,

    Elnagar, G. N., and Kazemi, M. A., “Pseudospectral Chebyshev optimal control of constrained nonlinear dynamical systems,” Computational Optimization and Applications, Vol. 11, 1998, pp. 195–217. https://doi.org/10.1023/A:1018694111831

    1998 doi

  41. [49]

    Path-constrained trajectory optimization using sparse sequential quadrtic programming,

    Betts, J. T., and Huffman, W. P., “Path-constrained trajectory optimization using sparse sequential quadrtic programming,” Journal of Guidance, Control, and Dynamics, Vol. 16, No. 1, 1993, pp. 59–68. https://doi.org/10.2514/3.11428

    1993 doi

  42. [50]

    Vehicles: Falcon 9,

    SpaceX, “Vehicles: Falcon 9,” , 2024. URL https://www.spacex.com/vehicles/falcon-9/

    2024

  43. [51]

    Autonomous Precision Landing of Space Rockets,

    Blackmore, L., “Autonomous Precision Landing of Space Rockets,”Frontiers of Engineering: Reports on Leading-Edge Engineering from the 2016 Symposium, Vol. 46, 2016, pp. 15–20

    2016

  44. [52]

    CVXGEN: a code generator for embedded convex optimization,

    Mattingley, J., and Boyd, S., “CVXGEN: a code generator for embedded convex optimization,”Optimization and Engineering, Vol. 13, 2011, pp. 1–27. https://doi.org/10.1007/s11081-011-9176-9

    2011 doi

  45. [53]

    Vehicles: Starship,

    SpaceX, “Vehicles: Starship,” , 2024. URL https://www.spacex.com/vehicles/starship/

    2024

  46. [54]

    Space Launch in 50 Years: Abundance at Last?

    Shotwell, G. E., and Blackmore, L., “Space Launch in 50 Years: Abundance at Last?”National Academy of Engineering, Vol. 46, 2016, pp. 144–146

    2016

  47. [55]

    SpaceX successfully lands a Starship test flight,

    Timmer, J., “SpaceX successfully lands a Starship test flight,” , 2021. URL https://arstechnica.com/science/2021/05/spacex- successfully-lands-a-starship-test-flight/

    2021

  48. [56]

    Customized Real-Time First-Order Methods for Onboard Dual Quaternion-based 6-DoF Powered-Descent Guidance,

    Kamath, A. G., Elango, P., Kim, T., Mceowen, S., Yu, Y., Carson, J. M., Mesbahi, M., and Acikmese, B., “Customized Real-Time First-Order Methods for Onboard Dual Quaternion-based 6-DoF Powered-Descent Guidance,” 2023. https: //doi.org/10.2514/6.2023-2003

    2023 doi

  49. [57]

    Real-Time Sequential Conic Optimization for Multi-Phase Rocket Landing Guidance,

    Kamath, A. G., Elango, P., Yu, Y., Mceowen, S., Chari, G. M., Carson III, J. M., and Açıkmeşe, B., “Real-Time Sequential Conic Optimization for Multi-Phase Rocket Landing Guidance,”IFAC-PapersOnLine, Vol. 56, No. 2, 2023, pp. 3118–3125. https://doi.org/https://doi.org/10.1016/...

    2023 doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.