Paved guideway topology optimization for pedestrian traffic under Nash equilibrium


Without proper flow channelization, congestion and overcrowding in pedestrian traffic may lead to significant inefficiency and safety hazards. Thus, the design of guideway networks that provide a fine balance between traffic congestion and infrastructure construction investment is vital. This paper presents a mathematical formulation and topology optimization framework for paved pedestrian guideway design under physics-based traffic equilibrium in a continuous space. Pedestrians are homogeneous, and their destination and path choices under the Nash equilibrium condition are described by a set of nonlinear partial differential equations. The design framework optimizes the deployment of pavement, which alters the road capacity and directly affects pedestrians’ free flow travel speed. A maximum crowd density constraint is included in the design model to address public safety concerns (e.g., over stampede risks). A series of numerical experiments are conducted to illustrate the effectiveness of the proposed model as well as solution techniques. The proposed framework, which builds on the traffic equilibrium theory, produces optimized guideway designs with controllable maximum pedestrian density, accounts for budget constraints (through an adjustable multiplier that balances pavement construction and travel costs), and allows for control of the spatial configuration of road branches. Comparison with lamellar structures and more conventional guideway designs demonstrates better performance of the outcomes from the proposed modeling and optimization framework.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13
Fig. 14
Fig. 15
Fig. 16
Fig. 17
Fig. 18


  1. 1.

    This function form is analogous to the well-known BPR function (U.S. Bureau of Public Roads), but other non-decreasing functions are also acceptable.

  2. 2.

    The choice of facility i depends on the pedestrian’s location x, i.e., i(x). However, we note that \(\boldsymbol {x} \in {\mathscr{A}}_{i}\) equivalently indicates the facility choice of the pedestrians at x. Hence, we simply use facility index i without the argument, for notation convenience.

  3. 3.

    Also called user equilibrium (UE) in the transportation literature, as originally described in Wardrop (1952).

  4. 4.

    Wadbro and Noreland (2019) used an interesting linear heat conduction model analogy to approach a similar problem; i.e. it assumes a linear relation between the local flux f(x) and the potential gradient ∇ϕ(x) with a constant conductivity factor κ. The model we introduced in (16)–(19), based on the Nash equilibrium condition of travelers in the continuous domain, shows that the κ in (17) should be a nonlinear function of |∇ϕ|, see (15).

  5. 5.

    The traffic equilibrium model is based on “macroscopic” fluid approximation, where pedestrians are described not as discrete particles but by continuous flux. Hence, the PDE and optimization model do not impose any requirement on the minimum thickness of guideway paths. The congestion delay is dictated by the ratio of flux intensity |f| to the capacity α, and hence the optimization model tends to provide capacity to the area where flux concentrates. When traffic is dispersed and relatively light (e.g., near the entrances and exits), it is reasonable for the optimization model to yield very thin roads. The density filter is hence needed as a regularization approach to control the length scale to avoid overly thin roads—for practical construction convenience and aesthetic purposes.



road capacity at location x

α 0 :

minimum road capacity

α :

road capacity of element

\(\alpha _{\max \limits }\) :

maximum road capacity

\(\bar {q}\) :

traffic inflow or outflow at Ω

β :

shadow price of construction

α :

discretized road capacity vector

λ g :

adjoint vector in sensitivity analysis of constraint function g

λ J :

adjoint vector in sensitivity analysis of objective function J

Ω :

two-dimensional space

ρ :

discretized pedestrian density vector

\(\boldsymbol {\tilde {F}}\) :

matrix storing the F of previous m AAR iterations

\(\boldsymbol {\tilde {R}}\) :

matrix storing the RF of previous m AAR iterations

F :

global flux magnitude vector

G :

matrix mapping global degree of freedom vector Φi to ∇ϕ at the center of element

I :

identity matrix

\(\boldsymbol {k}^{0}_{\ell }\) :

element stiffness matrix of element with κ = 1

K α :

global stiffness matrix

M :

matrix defined by \(\boldsymbol {M_{\ell }} := \boldsymbol {G_{\ell }}^{T} \boldsymbol {G_{\ell }}\)


unit normal vector at of Ω

P :

density filter matrix

Q :

global traffic inflow/outflow vector

R :

residual vector in Newton’s method

R F :

global residual vector in AAR method

v :

vector of element area

x i :

Location of facility i

z :

design variable vector

z i n i :

initial design variable vector

\(\boldsymbol {z}_{\max \limits }\) :

upper bounds of design variable

η :

Free flow travel time per unit distance

κ :

variable defined as \(\kappa :=\frac {\mid \mathbf {f}\mid }{c}\)

κ :

κ value associated with element , assumed to be constant inside element

\(\kappa _{\min \limits }\) :

a small value to void numerical singularities

f :

flux vector of pedestrian flows

\({\mathscr{A}}_{i}\) :

area from where pedestrian travels to facility i

\({\mathscr{H}}_{1}\) :

Sobolev space

\({\mathscr{U}}\) :

space of trial functions

\({\mathscr{U}}^{0}\) :

space of test functions

\({\mathscr{U}}_{h}\) :

finite-dimensional trial function space

ϕ :

total generalized cost

ϕ h :

finite-dimensional total generalized cost

Φ i :

value of ϕh at node i

ψ :

test function in weak form

ψ h :

finite-dimensional test function

Ψi :

values of ψh at node i

ρ :

pedestrian density

\(\rho _{\max \limits }\) :

upper bound for pedestrian density

τ o p t :

tolerance of change of design variable for terminating optimization

x :

position vector

𝜃 :

step size in AAR

b 1 :

monetary cost to travel a unit distance

b 2 :

ratio of capacity to free flow travel speed

c :

generalized cost to travel a unit distance

C R :

prorated unit cost for road construction

C T :

monetary value of one unit pedestrian time

E :

area of element


prorated cost per unit time for setting a capacity of α(x) for a unit area near x

g :

parameter characterizing travel cost sensitivity to congestion

h e :

width of a square quadrilateral element

J :

objective function

J R :

total construction cost

J T :

total transportation cost

M :

number of nodes in an finite element mesh

N i :

global Lagrange basis function associated with node i

\(N_{\max \limits }\) :

maximum number of optimization steps

p :

travel path

p n :

power in p-norm of a n-dimensional vector

P r :

period of applying Anderson extrapolation in AAR

q :

traffic inflow or outflow in Ω

R :

filter radius

r 0 :

radius of facilities

T h :

finite element partition


  1. Aage N, Andreassen E, Lazarov BS, Sigmund O (2017) Giga-voxel computational morphogenesis for structural design. Nature 550:84

    Google Scholar 

  2. Alexandersen J, Andreasen C (2020) A review of topology optimisation for fluid-based problems. Fluids 5(29)

  3. Alexandersen J, Sigmund O, Aage N (2016) Large scale three-dimensional topology optimisation of heat sinks cooled by natural convection. Int J Heat Mass Transfer 100:876–891

    Google Scholar 

  4. An S, Cui N, Bai Y, Xie W, Chen M, Ouyang Y (2015) Reliable emergency service facility location under facility disruption, en-route congestion and in-facility queuing. Transp Res E 82:199–216

    Google Scholar 

  5. Anderson DG (1965) Iterative procedures for nonlinear integral equations. J ACM (JACM) 12 (4):547–560

    MathSciNet  MATH  Google Scholar 

  6. Bai Y, Hwang T, Kang S, Ouyang Y (2011) Biofuel refinery location and supply chain planning under traffic congestion. Trans Res Part B Meth 45(1):162–175

    Google Scholar 

  7. Bai Y, Ouyang Y, Pang J (2016) Enhanced models and improved solution for competitive biofuel supply chain design under land use constraints. Eur J Oper Res 249(1):281–297

    MathSciNet  MATH  Google Scholar 

  8. Banerjee AS, Suryanarayana P, Pask JE (2016) Periodic pulay method for robust and efficient convergence acceleration of self-consistent field iterations. Chem Phys Lett 647:31–35

    Google Scholar 

  9. BBC News (2015) Shanghai new year crush kills 36. Web link:

  10. Beghini LL, Beghini A, Katz N, Baker WF, Paulino GH (2014) Connecting architecture and engineering through structural topology optimization. Eng Struct 59:716–726

    Google Scholar 

  11. Bendsøe MP, Kikuchi N (1988) Generating optimal topologies in structural design using a homogenization method. Comput Meth Appl Mechan Eng 71(2):197–224.

    MathSciNet  MATH  Google Scholar 

  12. Bendsøe MP, Sigmund O (2003) Topology optimization: theory, methods, and applications. Springer, Berlin

    Google Scholar 

  13. Borrvall T, Petersson J (2003) Topology optimization of fluids in stokes flow. Int J Numer Methods Fluids 41(1):77–107

    MathSciNet  MATH  Google Scholar 

  14. Bourdin B (2001) Filters in topology optimization. Int J Numer Methods Eng 50(9):2143–2158

    MathSciNet  MATH  Google Scholar 

  15. Christiansen RE, Wang F, Sigmund O (2019) Designing photonic topological insulators with quantum-spin-hall edge states using topology optimization. Nanophotonics 8:1363–1369

    Google Scholar 

  16. Clausen A, Wang F, Jensen JS, Sigmund O, Lewis JA (2015) Topology optimized architectures with programmable poisson’s ratio over large deformations. Adv Mater 27(37):5523–5527

    Google Scholar 

  17. Da D, Yvonnet J, Xia L, Li G (2018) Topology optimization of particle-matrix composites for optimal fracture resistance taking into account interfacial damage. Int J Numer Methods Eng 115(5):604–626

    Google Scholar 

  18. Deaton JD, Grandhi RV (2014) A survey of structural and multidisciplinary continuum topology optimization: post 2000. Struct Multidiscip Optim 49(1):1–38.

    MathSciNet  Article  Google Scholar 

  19. Duysinx P, Sigmund O (1998) New developments in handling stress constraints in optimal material distribution, 7th AIAA/USAF/NASA/ISSMO Symposium on Multidisciplinary Analysis and Optimization, American Institute of Aeronautics and Astronautics, 0.

  20. Gersborg-Hansen A, Bendsøe MP, Sigmund O (2006) Topology optimization of heat conduction problems using the finite volume method. Struct Multidiscip Optim 31(4):251–259.

    MathSciNet  Article  MATH  Google Scholar 

  21. Gladstone R (2015) Death toll from hajj stampede reaches 2,411 in new estimate

  22. Guenther R, Lee J (1996) Partial differential equations of mathematical physics and integral equations. Dover books on mathematics, Dover Publications

  23. Hajibabai L, Ouyang Y (2013) Integrated planning of supply chain networks and multimodal transportation infrastructure expansion: model development and application to the biofuel industry. Comput-Aided Civ Inf Eng 28(4):247–259

    Google Scholar 

  24. Hajibabai L, Bai Y, Ouyang Y (2014) Joint optimization of freight facility location and pavement infrastructure rehabilitation under network traffic equilibrium. Trans Res Part B Meth 63:38–52

    Google Scholar 

  25. Helbing D (1991) A mathematical model for the behavior of pedestrians. Behav Sci 36(4):298–310

    Google Scholar 

  26. Helbing D, Molnar P (1995) Social force model for pedestrian dynamics. Phys Rev E 51 (5):4282

    Google Scholar 

  27. Helbing D, Buzna L, Johansson A, Werner T (2005) Self-organized pedestrian crowd dynamics: experiments, simulations, and design solutions. Trans Sci 39(1):1–24

    Google Scholar 

  28. Helbing Dirk JK, Molnar P (1997) Modelling the evolution of human trail systems. Nature 388:47–50

    Google Scholar 

  29. Hoogendoorn S, Campanella M, Daamen W (2011) Fundamental Diagrams for Pedestrian Networks. In: Peacock R, Kuligowski E, Averill J (eds) Pedestrian and Evacuation Dynamics. Springer, Boston, MA, pp 255–264

  30. Jiang L, Li J, Shen C, Yang S, Han Z (2014) Obstacle optimization for panic flow-reducing the tangential momentum increases the escape speed. PloS one 9(12) e115:463

    Google Scholar 

  31. Johansson A, Helbing D (2007) Pedestrian flow optimization with a genetic algorithm based on boolean grids. Springer, pp 267–272

  32. Konur D, Geunes J (2011) Analysis of traffic congestion costs in a competitive supply chain. Trans Res Part E Logist Trans Rev 47(1):1–17

    Google Scholar 

  33. Konur D, Geunes J (2012) Competitive multi-facility location games with non-identical firms and convex traffic congestion costs. Trans Res Part E Logist Trans Rev 48(1):373–385

    Google Scholar 

  34. Li AC, Nozick L, Xu N, Davidson R (2012) Shelter location and transportation planning under hurricane conditions. Trans Res Part E Logist Trans Rev 48(4):715–729

    Google Scholar 

  35. Little JDC (1961) A proof for the queuing formula: L = w. Oper Res 9:296–435

    MathSciNet  Google Scholar 

  36. Maute K, Allen M (2004) Conceptual design of aeroelastic structures by topology optimization. Struct Multidiscip Optim 27(1):27–42.

    Article  Google Scholar 

  37. Ouyang Y, Wang Z , Yang H (2015) Facility location design under continuous traffic equilibrium. Transp Res Part B 81(1):18–33

    Google Scholar 

  38. Russ JB, Waisman H (2020) A novel topology optimization formulation for enhancing fracture resistance with a single quasi-brittle material. Int J Numer Methods Eng 121(13):2827–2856

    MathSciNet  Google Scholar 

  39. Ryu JC, Park FC, Kim YY (2012) Mobile robot path planning algorithm by equivalent conduction heat flow topology optimization. Struct Multidiscip Optim 45(5):703–715

    MathSciNet  MATH  Google Scholar 

  40. Sherali HD, Carter TB, Hobeika AG (1991) A location-allocation model and algorithm for evacuation planning under hurricane/ flood conditions. Trans Res Part B Meth 25(6):439–452

    Google Scholar 

  41. Sigmund O (1994) Materials with prescribed constitutive parameters: an inverse homogenization problem. Int J Solids Struct 31(17):2313–2329

    MathSciNet  MATH  Google Scholar 

  42. Sigmund O (1995) Tailoring materials with prescribed elastic properties. Mech Mater 20(4):351–368

    Google Scholar 

  43. Sigmund O, Maute K (2013) Topology optimization approaches. Struct Multidiscip Optim 48 (6):1031–1055.

    MathSciNet  Article  Google Scholar 

  44. Suryanarayana P, Pratapa PP, Pask JE (2019) Alternating anderson-richardson method: an efficient alternative to preconditioned krylov methods for large, sparse linear systems. Comput Phys Commun 234:278–285

    Google Scholar 

  45. Svanberg K (1987) The method of moving asymptotes – a new method for structural optimization. Int J Numer Methods Eng 24(2):359–373

    MathSciNet  MATH  Google Scholar 

  46. Wadbro E, Noreland D (2019) Continuous transportation as a material distribution topology optimization problem. Struct Multidiscip Optim 59(5):1471–1482

    MathSciNet  Google Scholar 

  47. Wang F, Sigmund O, Jensen JS (2014) Design of materials with prescribed nonlinear properties. Journal of the Mechanics and Physics of Solids 69:156–174

    MathSciNet  Google Scholar 

  48. Wang F, Christiansen RE, Yu Y, Mørk J, Sigmund O (2018) Maximizing the quality factor to mode volume ratio for ultra-small photonic crystal cavities. Appl Phys Lett 113(24):241, 101

    Google Scholar 

  49. Wang Z (2017) Planning Service Facilities and Infrastructures Under Continuous Traffic Equilibrium. PhD Dissertation. University of Illinois at Urbana-Champaign

  50. Wang Z, Ouyang Y (2016) On solving a class of continuous traffic equilibrium problems and planning facility location under congestion. Revision Under Review

  51. Wang Z, Xie S, Ouyang Y (2019) Planning facility location in a continuous space under congestion and disruption risks. Under Revision

  52. Wardrop J. (1952) Some Theoretical Aspects of Road Traffic Research. ICE Proceedings: Engineering Divisions. pp. 325–362.

  53. Xia L, Breitkopf P (2015) Design of materials using topology optimization and energy-based homogenization approach in matlab. Struct Multidiscip Optim 52(6):1229–1241.

    MathSciNet  Google Scholar 

  54. Yan S, Wang F, Sigmund O (2018) On the non-optimality of tree structures for heat conduction. Int J Heat Mass Trans 122:660–680 .

    Google Scholar 

  55. Yang H (1996) A spatial price equilibrium model with congestion effects. Ann Reg Sci 30 (4):359–371

    Google Scholar 

  56. Yang H, Wong S (2000) A continuous equilibrium model for estimating market areas of competitive facilities with elastic demand and market externality. Transp Sci 34(2):216–227

    MATH  Google Scholar 

  57. Yang H, Yagar S, Iida Y (1994) Traffic assignment in a congested discrete/continuous transportation system. Trans Res Part B Meth 28(2):161–174

    Google Scholar 

  58. Zhang X, Ramos AS Jr, Paulino GH (2017) Material nonlinear topology design using the ground structure method with a discrete filter scheme. Struct Multidiscip Optim 55(6):2045–2072

    MathSciNet  Google Scholar 

  59. Zhang XS, de Sturler E, Paulino GH (2017) Stochastic sampling for deterministic structural topology optimization with many load cases: Density-based and ground structure approaches. Comput Methods Appl Mech Eng 325:463–487

    MathSciNet  MATH  Google Scholar 

  60. Zhang XS, Paulino GH, Ramos AS Jr (2018) Multi-material topology optimization with multiple volume constraints: a ground structure approach involving material nonlinearity. Struct Multidiscip Optim 57:161–182

    MathSciNet  Google Scholar 

  61. Zhu JH, Zhang WH, Xia L (2016) Topology optimization in aircraft and aerospace structures design. Arch Comput Methods Eng 23(4):595–622.

    MathSciNet  Article  MATH  Google Scholar 

Download references


The authors would like to thank Dr. Ole Sigmund for suggesting Example 1. We would also like to acknowledge Dr. K. Svanberg for providing the MMA code.


The work by X. S. Zhang and W. Li were supported in part by the University of Illinois. The work by Y. Ouyang was supported in part by the U.S. National Science Foundation via Grant CMMI-1662825.

Author information



Corresponding author

Correspondence to Xiaojia Shelly Zhang.

Ethics declarations

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Replication of results

Data are available from authors upon request.

Responsible Editor: Mehmet Polat Saka


Appendix A: A hybrid solution scheme for the nonlinear state equations

To address the challenges described in Section 3.1 associated with solving the nonlinear state equation (29), we propose a robust hybrid solution strategy which combines AAR, a fixed-point iteration method proposed by Suryanarayana et al. (2019) and Banerjee et al. (2016), and the standard Newton’s method. In the proposed solution strategy, we first use the AAR to bring the iterate close to the solution and then apply the Newtwon’s method to achieve fast convergence.

In the AAR, in order to avoid the difficulty related to the initial guess Φ = 0, we introduce a new global residual vector as

$$ \boldsymbol{R}_{F}(\boldsymbol{\alpha},\boldsymbol{F})={\sum}_{\ell}^{n}\kappa_{\ell}(\alpha_{\ell},|f|_{\ell})|\nabla_{E_{\ell}}\phi_{h}(\boldsymbol{F})|-\boldsymbol{F}=\boldsymbol{0}, $$

where |f| denotes the flux in element , \(\boldsymbol {F}\in \mathbb {R}^{n\time 1}\) is a vector collecting all the element fluxes, namely, F = [|f|1,...,|f|n]T; and κ(α,|f|) is defined by relation (13). We note that, unlike the original residual vector R(α,Φ) defined in (30), where Φ is the independent vector, the new residual vector RF(α,F) uses the flux vector F as the independent vector and Φ(F) is obtained from Φ = (Kα(κ))− 1Q, where we recall that κ = [κ1,...,κn]T.

Although defined in different forms and having different independent variables, we can show that RF(α,F) = 0 and R(α,Φ) = 0 are in fact equivalent in the sense that solution of RF(α,F) = 0 is also a solution of R(α,Φ) = 0 and vice versa. The advantage of using RF(α,F) in the AAR is that it allows us to use a non-zero initial guess for vector F which, according to Fig. 1, will can lead to well-conditioned matrix Kα. In our implementation, we use F = 1 as the initial guess for AAR in the first optimization step and, in the subsequent optimization steps, F is initialized using the converged F from the previous optimization step. This choice has been shown to be effective and robust for all the numerical examples in this work.

The AAR iteration to update the flux vector F as:

$$ \boldsymbol{F}^{(k+1)} = \max\Big(\boldsymbol{F}^{(k)} + \mathbf{B}^{(k)} \boldsymbol{R}_{F}(\boldsymbol{\alpha},\boldsymbol{F}^{(k)}),\boldsymbol{0}\Big), $$

where F(k) is the flux vector at the k th AAR iteration, \(\max \limits (\cdot ,\cdot )\) stands for the element-wise maximum operator between the two vectors, and the matrix B(k) is defined by:

$$ \mathbf{B}^{(k)}{}={}\left\{{}\begin{array}{ll}{\theta \mathbf{I}} & {\text { if }(k{}+{}1) / Pr \notin \mathbb{N}} \\ {\theta \mathbf{I}-{}\left( {}\tilde{\mathbf{F}}^{(k)}{}+{}\theta \tilde{\mathbf{R}}^{(k)}{}\right){}\left( {}\tilde{\mathbf{R}}^{(k),T} \tilde{\mathbf{R}}^{(k)}{}\right)^{-1} \tilde{\mathbf{R}}^{(k),T}} & {\text { if }(k{}+{}1) / Pr \in \mathbb{N}}\end{array}\right. $$

where 𝜃 is the step size, Pr is the period of applying Anderson mixing (Anderson 1965), and \(\tilde {\mathbf {F}}^{(k)} \in \mathbb {R}^{n\times m}\) and \(\tilde {\mathbf {R}}^{(k)} \in \mathbb {R}^{n\times m}\) are matrices collecting history information of flux and residual vectors:

$$ \tilde{\mathbf{F}}^{(k)}=\begin{bmatrix} {\varDelta}\boldsymbol{F}^{(k-m)} & {\varDelta} \boldsymbol{F}^{(k-m+1)} & \ldots& {\varDelta} \boldsymbol{F}^{(k-1)} \end{bmatrix} $$
$$ \tilde{\mathbf{R}}^{(k)}=\begin{bmatrix} {\varDelta} \boldsymbol{R}_{F}(\boldsymbol{\alpha},\boldsymbol{F}^{(k-m)}) & {\varDelta} \boldsymbol{R}_{F}(\boldsymbol{\alpha},\boldsymbol{F}^{(k-m+1)}) & {\ldots} & {\varDelta} \boldsymbol{R}_{F}(\boldsymbol{\alpha},\boldsymbol{F}^{(k-1)}) \end{bmatrix}, $$

and ΔF(j) = F(j+ 1)F(j), \({\varDelta } \boldsymbol {R}_{F}(\boldsymbol {\alpha },\boldsymbol {F}^{(j)})=\boldsymbol {R}_{F}(\boldsymbol {\alpha },\boldsymbol {F}^{(j+1)})-\boldsymbol {R}_{F}(\boldsymbol {\alpha },\boldsymbol {F}^{(j)})\). We note that, by defining matrix B using (40), we essentially apply a quasi-Newton Anderson mixing (Anderson 1965) every Pr iterations. In other iterations, the simple Richardson iteration is used. For all the examples in this work, we use 𝜃 = 0.5, m = 5, and Rr = 4. The AAR iteration is terminated when the 2-norm of the residual vector RF(α,F) is below 10− 3. The corresponding Φ = (Kα(κ))− 1Q is then taken as the initial guess of the Newton’s method described below.

In the Newton’s method, we will switch back to the original residual vector R in (30). At iteration k of the Newton’s method, we linearize the above nonlinear system of equations as

$$ \boldsymbol{R}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})\approx\boldsymbol{R}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}}^{(k)})+\mathbf{K}^{\alpha}_{T}\Big(\boldsymbol{\kappa}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}}^{(k)})\Big){\varDelta}\boldsymbol{{\varPhi}}^{(k)}=\boldsymbol{0}, $$

where \(\mathbf {K}^{\alpha }_{T}\doteq \partial \boldsymbol {R}/\partial \boldsymbol {{\varPhi }}\) is the tangent stiffness matrix. Solving the linearized system gives

$$ {\varDelta}\boldsymbol{{\varPhi}}^{(k)}=\Big[\mathbf{K}^{\alpha}_{T}\Big(\boldsymbol{\kappa}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}}^{(k)})\Big)\Big]^{-1}\boldsymbol{R}(\boldsymbol{{\varPhi}}^{k}), $$

which leads to the recurrent update formula Φk+ 1 = Φk + ΔΦk until the 2-norm of the residual vector is below tolerance 10− 5.

A consistently linearized tangent stiffness matrix is essential to ensure the convergence of the Newton’s method. Thus, in the remainder of this appendix, a detailed derivation of the consistent tangent stiffness matrix \(\mathbf {K}^{\alpha }_{T}\) is provided.

By definition and using the chain rule, we have

$$ \begin{array}{@{}rcl@{}} \mathbf{K}^{\alpha}_{T}\Big(\boldsymbol{\kappa}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})\Big)&=&\frac{\partial \boldsymbol{R}}{\partial \boldsymbol{{\varPhi}}}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})\\ &=& \frac{\partial \mathbf{K}^{\alpha}}{\partial \boldsymbol{{\varPhi}}}\Big(\boldsymbol{\kappa}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})\Big) \boldsymbol{{\varPhi}} + \mathbf{K}^{\alpha}\Big(\boldsymbol{\kappa}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})\Big) \end{array} $$

To obtain an explicit expression of \(\mathbf {K}^{\alpha }_{T}\), we first compute the following local matrix as

$$ [\mathbf{k}^{{\varPhi}}_{\ell}]_{ij}=\frac{\partial \kappa_{\ell}}{\partial {\varPhi}_{j}} \Big({\sum}_{m} [\mathbf{k}^{0}_{\ell}]_{im}{\varPhi}_{m}\Big), $$

where the derivative κ/Φj can be computed as follows

$$ \frac{\partial \kappa_{\ell}}{\partial {\varPhi}_{j}} = \frac{\partial \kappa_{\ell}}{\partial |\nabla_{E_{\ell}} \phi_{h}|} \frac{\partial |\nabla_{E_{\ell}} \phi_{h}|}{\partial {\varPhi}_{j}} = \frac{\partial \kappa_{\ell}}{\partial |\nabla_{E_{\ell}} \phi_{h}|}\frac{{\sum}_{i}[\mathbf{M}_{\ell}]_{ji}{\varPhi}_{i}}{2\sqrt{\boldsymbol{{\varPhi}}^{T}\mathbf{M}_{\ell}\boldsymbol{{\varPhi}}}} $$

Since κ is defined implicitly through \(h(\kappa _{\ell },\alpha _{\ell },|\nabla _{E_{\ell }}\phi _{h}|)=0\), see (15) and Footnote 4, we can compute the derivative \(\partial \kappa _{\ell }/\partial |\nabla _{E_{\ell }}\phi _{h}|\) as

$$ \begin{array}{@{}rcl@{}} \frac{d h}{d |\nabla_{E_{\ell}}\phi_{h}|}&=&\frac{\partial h(\kappa_{\ell},\alpha_{\ell},|\nabla_{E_{\ell}}\phi_{h}|)}{\partial \kappa_{\ell}}\frac{\partial \kappa_{\ell}}{\partial |\nabla_{E_{\ell}}\phi_{h}|}\\ &&+\frac{\partial h(\kappa_{\ell},\alpha_{\ell},|\nabla_{E_{\ell}}\phi_{h}|)}{\partial |\nabla_{E_{\ell}}\phi_{h}|}=0, \end{array} $$

which gives

$$ \frac{\partial \kappa_{\ell}}{\partial |\nabla_{E_{\ell}}\phi_{h}|}=-\frac{\partial h(\kappa_{\ell},\alpha_{\ell},|\nabla_{E_{\ell}}\phi_{h}|)/\partial |\nabla_{E_{\ell}}\phi_{h}|}{\partial h(\kappa_{\ell},\alpha_{\ell},|\nabla_{E_{\ell}}\phi_{h}|)/\partial \kappa_{\ell}} $$

Finally, the global tangent stiffness matrix can be given by

$$ \mathbf{K}^{\alpha}_{T}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})={\sum}_{\ell=1}^{M}\Big(\mathbf{k}^{{\varPhi}}_{\ell} + \kappa_{\ell}(\alpha_{\ell},|\nabla_{E_{\ell}}\phi_{h}|)\mathbf{k}^{0}_{\ell}\Big) $$

Appendix B: Sensitivity analysis

The sensitivities of the objective function J and constraint function g with respect to the design variable z can be obtained from the adjoint method as

$$ \begin{array}{@{}rcl@{}} &&{}\frac{\partial J}{\partial z_{\ell}}={\sum}_{m=1}^{n}\Big[\frac{\partial J(\boldsymbol{\alpha},\boldsymbol{\rho})}{\partial \alpha_{m}}+\frac{\partial J(\boldsymbol{\alpha},\boldsymbol{\rho})}{\partial \rho_{m}}\frac{\partial \rho_{m}}{\partial \alpha_{m}}+\boldsymbol{\lambda}_{J}^{T}\frac{\partial \boldsymbol{R}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})}{\partial \alpha_{m}}\Big]\frac{\partial \alpha_{m}}{\partial z_{\ell}}\quad \text{ and} \end{array} $$
$$ \begin{array}{@{}rcl@{}} &&{}\frac{\partial g}{\partial z_{\ell}}={\sum}_{m=1}^{n}\Big[\frac{\partial g(\boldsymbol{\rho})}{\partial \rho_{m}}\frac{\partial \rho_{m}}{\partial \alpha_{m}} +\boldsymbol{\lambda}_{g}^{T}\frac{\partial \boldsymbol{R}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})}{\partial \alpha_{m}}\Big]\frac{\partial \alpha_{m}}{\partial z_{\ell}} \end{array} $$

respectively, where λJ and λg are the vectors of adjoint variables given by

$$ \begin{array}{@{}rcl@{}} &&\boldsymbol{\lambda}_{J}=-\Big(\mathbf{K}^{\alpha}_{T}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})\Big)^{-T}\Big({\sum}_{m=1}^{n}\frac{\partial J(\boldsymbol{\alpha},\boldsymbol{\rho})}{\partial \rho_{m}}\frac{\partial \rho_{m}}{\partial \boldsymbol{{\varPhi}}}\Big)\quad\text{ and} \end{array} $$
$$ \begin{array}{@{}rcl@{}} &&\boldsymbol{\lambda}_{g}=-\Big(\mathbf{K}^{\alpha}_{T}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})\Big)^{-T}\Big({\sum}_{m=1}^{n}\frac{\partial g(\boldsymbol{\rho})}{\partial \rho_{m}}\frac{\partial\rho_{m}}{\partial \boldsymbol{{\varPhi}}}\Big) \end{array} $$

respectively, with \(\mathbf {K}^{\alpha }_{T}\) being evaluated at the converged solution of each optimization step and blackαm/z = [P]m.

In the above expressions for sensitivity analysis, the detailed expressions of J/αm, J/ρm, g/ρm and R/αm are given below:

$$ \begin{array}{@{}rcl@{}} \frac{\partial J(\boldsymbol{\alpha},\boldsymbol{\rho})}{\partial \alpha_{m}}&=&\beta C_{B} v_{m}; \quad\quad\frac{\partial J(\boldsymbol{\alpha},\boldsymbol{\rho})}{\partial \rho_{m}}=(1-\beta) C_{T} v_{m}; \end{array} $$
$$ \begin{array}{@{}rcl@{}} \frac{\partial g(\boldsymbol{\rho})}{\partial \rho_{m}}&=& \Big({\sum}_{j=1}^{n}\Big(\rho_{j}\Big)^{{p_{n}}}\Big)^{\frac{1}{{p_{n}}}-1}\Big(\rho_{m}\Big)^{{p_{n}}-1}; \end{array} $$
$$ \begin{array}{@{}rcl@{}} \frac{\partial \boldsymbol{R}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})}{\partial \alpha_{m}}&=&\frac{\partial \mathbf{K}^{\alpha}(\boldsymbol{\alpha},\boldsymbol{{\varPhi}})}{\partial \alpha_{m}}\boldsymbol{{\varPhi}}=\frac{\partial \kappa_{m}}{\partial \alpha_{m}}\mathbf{k}^{0}_{m}\boldsymbol{{\varPhi}} \end{array} $$

Additionally, the detailed expressions of κm/αm in the above expressions can be obtained in the similar manner as \(\partial \kappa _{m}/\partial |\nabla _{E_{m}}\phi _{h}|\) (i.e., (48)–(49)) as

$$ \frac{\partial \kappa_{m}}{\partial \alpha_{m}}=-\frac{\partial h(\kappa_{m},\alpha_{m},|\nabla_{E_{m}}\phi_{h}|)/\partial \alpha_{m}}{\partial h(\kappa_{m},\alpha_{m},|\nabla_{E_{m}}\phi_{h}|)/\partial \kappa_{m}}, $$

with κm and \(|\nabla _{E_{m}}\phi _{h}|\) being evaluated at the converged solution of each optimization step. Once κm/αm is obtained, we can further compute ρm/αm and ρm/Φ based on (34) as

$$ \begin{array}{@{}rcl@{}} \frac{\partial \rho_{m}(\alpha_{m},\boldsymbol{{\varPhi}})}{\partial \alpha_{m}}&=&\frac{\partial \rho_{m}}{\partial \alpha_{m}}+\frac{\partial \rho_{m}}{\partial \kappa_{m}}\frac{\partial \kappa_{m}}{\partial \alpha_{m}}\\ &=&-|\nabla_{E_{m}}\phi_{h}|\frac{\kappa_{m}}{\alpha_{m}}\Bigg[\frac{b_{2}}{\alpha_{m}}+g\Big(\frac{\kappa_{m}|\nabla_{E_{m}}\phi_{h}|}{\alpha_{m}^{g+1}}\Big)^{g}\Bigg]\\ &&+ |\nabla_{E_{m}}\phi_{h}|\Bigg[\frac{b_{2}}{\alpha_{m}}+(g+1)\Big(\frac{\kappa_{m}|\nabla_{E_{m}}\phi_{h}|}{\alpha_{m}}\Big)^{g}\Bigg]\frac{\partial \kappa_{m}}{\partial \alpha_{m}} \end{array} $$


$$ \begin{array}{@{}rcl@{}} \frac{\partial \rho_{m}(\alpha_{m},\boldsymbol{{\varPhi}})}{\partial \boldsymbol{{\varPhi}}}&=&\Bigg[\frac{\partial \rho_{m}}{\partial |\nabla_{E_{m}}\phi_{h}|}+\frac{\partial \rho_{m}}{\partial \kappa_{m}}\frac{\partial \kappa_{m}}{\partial |\nabla_{E_{m}}\phi_{h}|}\Bigg]\frac{\partial |\nabla_{E_{m}}\phi_{h}|}{\partial\boldsymbol{{\varPhi}}}\\ &=&\Big(\kappa_{m}+|\nabla_{E_{m}}\phi_{h}|\frac{\partial \kappa_{m}}{\partial |\nabla_{E_{m}}\phi_{h}|}\Big)\\ &&\times\Bigg[\frac{b_{2}}{\alpha_{m}}+(g+1)\Big(\frac{\kappa_{m}|\nabla_{E_{m}}\phi_{h}|}{\alpha_{m}}\Big)^{g}\Bigg]\frac{\mathbf{M}_{m}\boldsymbol{{\varPhi}}}{\sqrt{\boldsymbol{{\varPhi}}^{T}\mathbf{M}_{m}\boldsymbol{{\varPhi}}}}, \end{array} $$

where both κ/αm and \(\partial \kappa /\partial |\nabla _{E_{m}}\phi _{h}|\) are obtained by evaluating (58) and (49) at the converged solution of each optimization step.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zhang, X.S., Li, W. & Ouyang, Y. Paved guideway topology optimization for pedestrian traffic under Nash equilibrium. Struct Multidisc Optim 63, 1405–1426 (2021).

Download citation


  • Topology optimization
  • Guideway network design
  • Continuous traffic equilibrium
  • Transportation engineering
  • Nash equilibirum