Periodic solutions of a generalized Sitnikov problem


In this paper, we study the problem of determining whether a global family of even periodic solutions of a generalized Sitnikov problem, which emerges from a circular generalized Sitnikov problem, continues for all values of eccentricity in [0, 1) or ends in the equilibrium.

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

Fig. 1
Fig. 2


  1. Alekseev, V.M.: Quasirandom dynamical systems. II. One-dimensional nonlinear oscillations in a field with periodic perturbation. Math. USSR-Sbornik 6(4), 505–560 (1968)

    Article  Google Scholar 

  2. Alfaro, J.M., Chiralt, C.: Invariant rotational curves in Sitnikov’s problem. Celest. Mech. Dyn. Astron. 55(4), 351–367 (1993)

    ADS  MathSciNet  Article  Google Scholar 

  3. Belbruno, E., Llibre, J., Ollé, M.: On the families of periodic orbits which bifurcate from the circular Sitnikov motions. Celest. Mech. Dyn. Astron. 60(1), 99–129 (1994)

    ADS  MathSciNet  Article  Google Scholar 

  4. Beltritti, G., Mazzone, F., Oviedo, M.: The Sitnikov problem for several primary bodies configurations. Celest. Mech. Dyn. Astron. 130(7), 45 (2018)

    ADS  MathSciNet  Article  Google Scholar 

  5. Bountis, T., Papadakis, K.: The stability of vertical motion in the n-body circular Sitnikov problem. Celest. Mech. Dyn. Astron. 104(1), 205–225 (2009)

    ADS  MathSciNet  Article  Google Scholar 

  6. Corbera, M., Llibre, J.: Periodic orbits of the Sitnikov problem via a Poincaré map. Celest. Mech. Dyn. Astron. 77(4), 273–303 (2000)

    ADS  Article  Google Scholar 

  7. Corbera, M., Llibre, J.: On symmetric periodic orbits of the elliptic Sitnikov problem via the analytic continuation method. Contemp. Math. 292, 91–128 (2002)

    MathSciNet  Article  Google Scholar 

  8. Dvorak, R.: Numerical results to the Sitnikov-problem. Celest. Mech. Dyn. Astron. 56(1–2), 71–80 (1993)

    ADS  Article  Google Scholar 

  9. Hartman, P.: Ordinary differential equations, Classics in Applied Mathematics, vol. 38. Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA (2002). Corrected reprint of the second (1982) edition

  10. Jiménez-Lara, L., Escalona-Buendía, A.: Symmetries and bifurcations in the Sitnikov problem. Celest. Mech. Dyn. Astron. 79(2), 97–117 (2001)

    ADS  MathSciNet  Article  Google Scholar 

  11. Llibre, J., Ortega, R.: On the families of periodic orbits of the Sitnikov problem. SIAM J. Appl. Dyn. Syst. 7(2), 561–576 (2008)

    ADS  MathSciNet  Article  Google Scholar 

  12. Llibre, J., Simó, C.: Estudio cualitativo del problema de Sitnikov. Publicacions de la Secció de Matemàtiques, pp. 49–71(1980)

  13. Llibre, J., Moeckel, R., Simó, C.: Central configurations, periodic orbits, and Hamiltonian systems. Advanced Courses in Mathematics - CRM Barcelona. Birkhauser (2015)

  14. Magnus, W., Winkler, S.: Hill’s Equation. Courier Corporation, Chelmsford (2013)

    Google Scholar 

  15. Marchesin, M., Vidal, C.: Spatial restricted rhomboidal five-body problem and horizontal stability of its periodic solutions. Celest. Mech. Dyn. Astron. 115(3), 261–279 (2013)

    ADS  MathSciNet  Article  Google Scholar 

  16. Moser, J.: Stable and Random Motions in Dynamical Systems: With Special Emphasis on Celestial Mechanics, Annals Mathematics Studies, vol. 1. Princeton University Press, Princeton (1973)

    Google Scholar 

  17. Murray, C.D., Dermott, S.F.: Solar System Dynamics. Cambridge University Press, Cambridge (1999)

    Google Scholar 

  18. Ortega, R.: Symmetric periodic solutions in the Sitnikov problem. Archiv der Mathematik 107(4), 405–412 (2016)

    MathSciNet  Article  Google Scholar 

  19. Ortega, R., Rivera, A.: Global bifurcations from the center of mass in the Sitnikov problem. Discrete Contin. Dyn. Syst. B 14(2), 719–732 (2010)

    MathSciNet  Article  Google Scholar 

  20. Pandey, L.P., Ahmad, I.: Periodic orbits and bifurcations in the Sitnikov four-body problem when all primaries are oblate. Astrophys. Space Sci. 345(1), 73–83 (2013)

    ADS  Article  Google Scholar 

  21. Perdios, E., Markellos, V.V.: Stability and bifurcations of Sitnikov motions. Celest. Mech. Dyn. Astron. 42(1–4), 187–200 (1987)

    MathSciNet  Article  Google Scholar 

  22. Rabinowitz, P.H.: Some global results for nonlinear eigenvalue problems. J. Funct. Anal. 7(3), 487–513 (1971)

    MathSciNet  Article  Google Scholar 

  23. Rivera, A.: Periodic solutions in the generalized Sitnikov \((n+1)\)-body problem. SIAM J. Appl. Dyn. Syst. 12(3), 1515–1540 (2013)

    MathSciNet  Article  Google Scholar 

  24. Rivera, A., Andrés, M.: Bifurcación de soluciones periódicas en el problema de Sitnikov. Universidad de Granada, Granada (2012)

    Google Scholar 

  25. Sitnikov, K.: The existence of oscillatory motions in the three-body problem. Dokl. Akad. Nauk SSSR 133, 303–306 (1960)

    MathSciNet  MATH  Google Scholar 

  26. Soulis, P.S., Papadakis, K.E., Bountis, T.: Periodic orbits and bifurcations in the Sitnikov four-body problem. Celest. Mech. Dyn. Astron. 100(4), 251–266 (2008)

    ADS  MathSciNet  Article  Google Scholar 

  27. Zanini, C.: Rotation numbers, eigenvalues, and the Poincaré–Birkhoff theorem. J. Math. Anal. Appl. 279(1), 290–307 (2003)

    MathSciNet  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Gastón Beltritti.

Additional information

Publisher's Note

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



In this section, we enunciate and demonstrate some technical lemmas, which deal with Taylor series of \(\sqrt{q_\lambda (u,e)}\) and \(\int _{0}^{\pi }\sqrt{q_\lambda (u,e)}\mathrm{d}u\) and are used to demonstrate other results of the article.

Before enunciating the first lemma, it should be noted the definition of combinatory number, \(\begin{pmatrix} \gamma \\ n \end{pmatrix}:=\frac{\prod _{i=1}^{n}(\gamma -i+1)}{n!}\), where \(\gamma \in {\mathbb {R}}\) and \(n\in {\mathbb {N}}\).

Lemma 6

Let \(q_\lambda (u,e)\) be the function defined in (16). Then, for \(u \in (0,\pi )\), we have that

$$\begin{aligned} \sqrt{q_\lambda (u,e)}=\sqrt{8\lambda }\sum _{l=0}^{\infty }c_{\lambda ,l}(u)e^{l}, \end{aligned}$$


$$\begin{aligned}&c_{\lambda ,l}(u)=\sum _{j=0}^{\left[ \frac{l}{2}\right] }\sum _{k=0}^{l-2j}\begin{pmatrix} \frac{1}{2}-j\\ k \end{pmatrix}b_\lambda ^k(u)\cos ^{l-k-2j}(u)\begin{pmatrix} \frac{1}{2}\\ j \end{pmatrix}a_\lambda ^{j}(u), \end{aligned}$$

with \(a_\lambda (u)=\frac{\cos ^2(u)-3}{32\lambda }\), \(b_\lambda (u)=-\frac{32\lambda -2}{32\lambda }\cos (u)\).


Let us first observe that

$$\begin{aligned} \sqrt{q_\lambda (u,e)}= & {} \frac{1}{2}\frac{\sqrt{32\lambda -(32\lambda -2)\cos (u)e+(\cos ^2(u)-3)e^2}}{1-e\cos (u)}\\= & {} \sqrt{8\lambda }\frac{\sqrt{1+b_\lambda (u)e+a_\lambda (u)e^2}}{1-e\cos (u)} \end{aligned}$$

where \(a_\lambda (u)=\frac{\cos ^2(u)-3}{32\lambda }\) and \(b_\lambda (u)=-\frac{32\lambda -2}{32\lambda }\cos (u)\). As \(\sqrt{1+x}=\sum _{i=0}^{\infty }\begin{pmatrix} \frac{1}{2}\\ i \end{pmatrix}x^i\) for all \(-1<x< 1\), we have that

$$\begin{aligned} \sqrt{q_\lambda (u,e)}=\sqrt{8\lambda }f_\lambda (u,e)\sum _{n=0}^{\infty }\cos ^n(u)e^n, \end{aligned}$$

where \(f_\lambda (u,e):=\sum _{n=0}^{\infty }\begin{pmatrix} \frac{1}{2}\\ n \end{pmatrix}e^n(b_\lambda (u)+ea_\lambda (u))^n\).

Since \(\sqrt{1+x}\) is analytic in \((-1,1)\) and \(b_\lambda (u)e+a_\lambda (u)e^2\) is analytic, as a function of e in \({\mathbb {R}}\), we have that, for every \(u\in (0,\pi )\), \(\sqrt{1+b_\lambda (u)e+a_\lambda (u)e^2}=f_\lambda (u,e)\) is analytic, as a function of e, in \((-1,1)\). Then, for \(e\in [0,1)\)

$$\begin{aligned} f_\lambda (u,e)=\sum _{n=0}^{\infty }\frac{\partial ^nf_\lambda (u,0)}{\partial e^n}\frac{e^n}{n!}. \end{aligned}$$

Let us calculate \(\frac{\partial ^nf_\lambda (u,0)}{\partial e^n}\).

$$\begin{aligned} \frac{\partial ^nf_\lambda (u,0)}{\partial e^n}= & {} \sum _{m=\left[ \frac{n+1}{2}\right] }^{n}\begin{pmatrix} \frac{1}{2}\\ m \end{pmatrix}\frac{\partial ^n}{\partial e^n}(e^m(b_\lambda (u)+ea_\lambda (u))^m)\Big |_{e=0}\\= & {} \sum _{m=\left[ \frac{n+1}{2}\right] }^{n}\begin{pmatrix} \frac{1}{2}\\ m \end{pmatrix}\sum _{i=0}^{n} \begin{pmatrix} n\\ i \end{pmatrix}\left( \frac{\partial ^ie^m}{\partial e^i}\frac{\partial ^{n-i}}{\partial e^{n-i}}(b_\lambda (u)+ea_\lambda (u))^m\right) \bigg |_{e=0}\\= & {} \sum _{m=\left[ \frac{n+1}{2}\right] }^{n}\begin{pmatrix} \frac{1}{2}\\ m \end{pmatrix} \begin{pmatrix} n\\ m \end{pmatrix}\left( \frac{\partial ^me^m}{\partial e^m}\frac{\partial ^{n-m}}{\partial e^{n-m}}(b_\lambda (u)+ea_\lambda (u))^m\right) \bigg |_{e=0}\\= & {} \sum _{m=\left[ \frac{n+1}{2}\right] }^{n}\begin{pmatrix} \frac{1}{2}\\ m \end{pmatrix} \begin{pmatrix} n\\ m \end{pmatrix}\frac{(m!)^2}{(2m-n)!}b_\lambda ^{2m-n}(u)a_\lambda ^{n-m}(u). \end{aligned}$$


$$\begin{aligned} \sqrt{q_\lambda (u,e)}=\sqrt{8\lambda }\sum _{l=0}^{\infty }c_{\lambda ,l}(u)e^l, \end{aligned}$$


$$\begin{aligned} c_{\lambda ,l}(u)=\sum _{n=0}^{l}\frac{\partial ^nf_\lambda (u,0)}{\partial e^n}\frac{\cos ^{l-n}(u)}{n!}. \end{aligned}$$

Considering \(j=n-m\), and \(k=n-2j\), we have that

$$\begin{aligned} c_{\lambda ,l}(u)= & {} \sum _{n=0}^{l}\sum _{m=\left[ \frac{n+1}{2}\right] }^{n}\begin{pmatrix} \frac{1}{2}\\ m \end{pmatrix}\begin{pmatrix} n\\ m \end{pmatrix}\frac{(m!)^2}{(2m-n)!}\frac{b_\lambda ^{2m-n}(u)a_\lambda ^{n-m}(u)}{n!}\cos ^{l-n}(u)\nonumber \\= & {} \sum _{n=0}^{l}\sum _{m=\left[ \frac{n+1}{2}\right] }^{n} \left[ \prod _{s=1}^{m}\left( \frac{1}{2}-s+1\right) \right] \frac{b_\lambda ^{2m-n}(u)}{(2m-n)!}\frac{a_\lambda ^{n-m}(u)}{(n-m)!}\cos ^{l-n}(u)\nonumber \\= & {} \sum _{j=0}^{\left[ \frac{l}{2}\right] }\sum _{n=2j}^{l}\prod _{s=1}^{n-j}\left( \frac{1}{2}-s+1\right) \frac{b_\lambda ^{n-2j}(u)}{(n-2j)!}\cos ^{l-n}(u)\frac{a_\lambda ^j(u)}{j!}\nonumber \\= & {} \sum _{j=0}^{\left[ \frac{l}{2}\right] }\sum _{k=0}^{l-2j}\prod _{s=j+1}^{k+j}\left( \frac{1}{2}-s+1\right) \frac{b_\lambda ^k(u)}{k!}\cos ^{l-k-2j}(u)\prod _{i=1}^{j}\left( \frac{1}{2}-i+1\right) \frac{a_\lambda ^{j}(u)}{j!}\nonumber \\= & {} \sum _{j=0}^{\left[ \frac{l}{2}\right] }\sum _{k=0}^{l-2j}\begin{pmatrix} \frac{1}{2}-j\\ k \end{pmatrix}b_\lambda ^k(u)\cos ^{l-k-2j}(u)\begin{pmatrix} \frac{1}{2}\\ j \end{pmatrix}a_\lambda ^{j}(u). \end{aligned}$$

\(\square \)

Lemma 7

Let \(q_\lambda (u,e)\) be the function defined in (16). Then

$$\begin{aligned} \int _{0}^{\pi }\sqrt{q_\lambda (u,e)}\mathrm{d}u=\sqrt{8\lambda }\sum _{\tau =0}^{\infty }\int _{0}^{\pi }c_{\lambda ,2\tau }(u)\mathrm{d}ue^{2\tau }. \end{aligned}$$


Note first that \(\begin{pmatrix} \frac{1}{2}\\ k \end{pmatrix}(-1)^k<0\) if \(k\ge 1\), and \(\begin{pmatrix} \frac{1}{2}-j\\ k \end{pmatrix}(-1)^k>0\) if \(j\ge 1\) and \(k\ge 0\). Since the function \(-1<\frac{a_\lambda (u)}{1+b_\lambda (u)} < 0\) for \(u\in (0,\pi )\), and \(\sum _{k=0}^{\infty }\begin{pmatrix} \gamma \\ k \end{pmatrix}\beta ^k=(1+\beta )^{\gamma }\) for \(\gamma \in {\mathbb {R}}\) and \(|\beta |<1\), if \(u\in (0,\pi )\), we have that

$$\begin{aligned}&|c_{\lambda ,l}(u)|\\&\quad \le 1+ \sum _{k=1}^{l}\left| \begin{pmatrix} \frac{1}{2}\\ k \end{pmatrix}b_\lambda ^k(u)\right| \left| \cos ^{l-k}(u)\right| +\sum _{j=1}^{\left[ \frac{l}{2}\right] }\sum _{k=0}^{l-2j}\left| \begin{pmatrix} \frac{1}{2}-j\\ k \end{pmatrix}b_\lambda ^k(u)\right| \left| \cos ^{l-k-2j}(u)\begin{pmatrix} \frac{1}{2}\\ j \end{pmatrix}a_\lambda ^{j}(u)\right| \\&\quad \le 1-\sum _{k=1}^{\infty }\begin{pmatrix} \frac{1}{2}\\ k \end{pmatrix}b_\lambda ^k(u)-\sum _{j=1}^{\infty }\sum _{k=0}^{\infty }\begin{pmatrix} \frac{1}{2}-j\\ k \end{pmatrix}b_\lambda ^k(u)\begin{pmatrix} \frac{1}{2}\\ j \end{pmatrix}a_\lambda ^{j}(u)\\&\quad =2-\sqrt{1+b_\lambda (u)}\sum _{j=0}^{\infty }\begin{pmatrix} \frac{1}{2}\\ j \end{pmatrix}\left( \frac{a_\lambda (u)}{1+b_\lambda (u)}\right) ^j\\&\quad \le 2. \end{aligned}$$

Thus, by the dominated convergence theorem we have that

$$\begin{aligned} \int _{0}^{\pi }\sqrt{q_\lambda (u,e)}\mathrm{d}u=\sqrt{8\lambda }\sum _{l=0}^{\infty }\int _{0}^{\pi }c_{\lambda ,l}(u)\mathrm{d}ue^l, \end{aligned}$$

for every \(0\le e<1\). Taking into account the expression of \(c_{\lambda ,l}(u)\), if l is an odd number then \(\int _{0}^{\pi }c_{\lambda ,l}(u)\mathrm{d}u=0\). Therefore,

$$\begin{aligned} \int _{0}^{\pi }\sqrt{q_\lambda (u,e)}\mathrm{d}u=\sqrt{8\lambda }\sum _{\tau =0}^{\infty }\int _{0}^{\pi }c_{\lambda ,2\tau }(u)\mathrm{d}ue^{2\tau }. \end{aligned}$$

\(\square \)

Remark 5

From (33) in Lemma 6, if we define \(h_\tau (u,\omega ):=\cos ^{2\tau -2\omega }(u)(3-\cos ^2(u))^\omega \), we have that

$$\begin{aligned}&c_{\lambda ,2\tau }(u) =\sum _{n=0}^{2\tau }\sum _{m=\left[ \frac{n+1}{2}\right] }^{n}\begin{pmatrix} \frac{1}{2}\\ m \end{pmatrix}\begin{pmatrix} n\\ m \end{pmatrix}\frac{(m!)^2}{(2m-n)!}\frac{b_\lambda ^{2m-n}(u)a_\lambda ^{n-m}(u)}{n!}\cos ^{2\tau -n}(u)\\&\quad =\sum _{n=0}^{2\tau }\sum _{m=\left[ \frac{n+1}{2}\right] }^{n}\left[ \prod _{s=1}^{m}\left( \frac{1}{2}-s+1\right) \right] \frac{(-1)^m}{(n-m)!}\frac{1}{(2m-n)!}\frac{(32\lambda -2)^{2m-n}}{(32\lambda )^m} h_{\tau }(u,n-m). \end{aligned}$$

Now, taking into account that \(\int _{0}^{\pi }\cos ^i(x)\mathrm{d}x=\frac{i-1}{i}\int _{0}^{\pi }\cos ^{i-2}(x)\mathrm{d}x\) for \(i\ge 2\), if \(\tau \ge n-m\ge 0\) then

$$\begin{aligned}&\int _{0}^{\pi } h_{\tau }(u,n-m)\mathrm{d}u\\&\quad =\sum _{i=0}^{n-m}\begin{pmatrix} n-m\\ i \end{pmatrix}3^{i}(-1)^{n-m-i}\int _{0}^{\pi }\cos ^{2(\tau -i)}(u)\mathrm{d}u\\&\quad =\sum _{i=0}^{n-m}\begin{pmatrix} n-m\\ i \end{pmatrix}3^{i}(-1)^{n-m-i}\prod _{j=1}^{\tau -i}\frac{2\tau -2i-2j+1}{2\tau -2i-2j+2}\pi \\&\quad =:{\bar{g}}(\tau ,n-m)\pi . \end{aligned}$$


$$\begin{aligned}&\int _{0}^{\pi }c_{\lambda ,2\tau }(u)\mathrm{d}u\nonumber \\= & {} \sum _{n=0}^{2\tau }\sum _{m=\left[ \frac{n+1}{2}\right] }^{n}\left[ \prod _{s=1}^{m}\left( \frac{1}{2}-s+1\right) \right] \frac{(-1)^m}{(n-m)!}\frac{1}{(2m-n)!}\frac{(32\lambda -2)^{2m-n}}{(32\lambda )^m}{\bar{g}}(\tau ,n-m)\pi \nonumber \\= & {} : g_\lambda (\tau )\pi . \end{aligned}$$

Note that \(\int _{0}^{\pi }c_{\lambda ,2\tau }(u)\mathrm{d}u\) is always a product between a number \(g_\lambda (\tau )\) and \(\pi \), so we can calculate with a symbolic computation software the expression of \(g_\lambda (\tau )\) without error. Thus, for example, we can verify that

$$\begin{aligned}&\int _{0}^{\pi }c_{\lambda ,0}(u)\mathrm{d}u=\pi \\&\quad \int _{0}^{\pi }c_{\lambda ,2}(u)\mathrm{d}u=\frac{768 \lambda ^{2} - 64 \lambda - 1}{4096 \lambda ^{2}}\pi \\&\quad \int _{0}^{\pi }c_{\lambda ,4}(u)\mathrm{d}u=\frac{3 \left( 2293760 \lambda ^{4} - 163840 \lambda ^{3} - 4608 \lambda ^{2} - 128 \lambda - 5\right) }{67108864 \lambda ^{4}}\pi . \end{aligned}$$

In general it seems to be that \(\int _{0}^{\pi }c_{\lambda ,2\tau }(u)\mathrm{d}u>0\) for all \(\lambda \ge \frac{1}{8}\) and \(\tau \in {\mathbb {N}}_0\).

The code used in the Python library “Sympy” with which we calculate \(\int _{0}^{\pi }c_{\lambda ,2\tau }(u)\mathrm{d}u\) is described below

$$\begin{aligned}&\text {def conv(m,n):}\\&\quad \text {return factorial(m)/factorial(m-n)/factorial(n)}\\&\quad \text { def intcos(s):}\\&\quad \text {return prod(Rational(2*s-2*j+1,2*s-2*j+2) for j in range(1,s+1))}\\&\quad \text {def bargt(t,s):}\\&\quad \text {return sum((-1)**(s-i)*3**(i)*conv(s,i)*intcos(t-i)\ for i in}\\&\quad \text {range(s+1))}\\&\quad \text {def gt(tau,lam):}\\&\quad \text {A=sum(sum((-1)**m*prod(Rational(1,2)-s+1}\\&\quad \text {for s in range(1,m+1))*1/factorial(n-m)/factorial(2*m-n)}\\&\quad \text {*(32*lam-2)**(2*m-n)/(32*lam)**m*bargt(tau,n-m) for m in}\\&\quad \text {range(floor(Rational(n+1,2)),n+1)) for n in range(2*tau+1))}\\&\quad \text {return A} \end{aligned}$$

The function \(\text {gt(tau,lam)}\) calculates \(g_\lambda (\tau )\) with error zero.

Lemma 8

Let \(c_{1,l}(u)\) be the coefficient defined in (32). Then

$$\begin{aligned} \int _{0}^{\pi }c_{1,2\tau }(u)\mathrm{d}u\ge -\frac{\pi }{4}. \end{aligned}$$

for all \(\tau \in {\mathbb {N}}_0\).


For \(u\in (0,\frac{\pi }{2})\), we have that

$$\begin{aligned}&c_{1,2\tau }(u)\nonumber \\&\quad =\sum _{k=0}^{2\tau }\begin{pmatrix} \frac{1}{2}\\ k \end{pmatrix}b_1^k(u)\cos ^{2\tau -k}(u)+\sum _{j=1}^{\tau }\sum _{k=0}^{2\tau -2j}\begin{pmatrix} \frac{1}{2}-j\\ k \end{pmatrix}b_1^k(u)\cos ^{2\tau -k-2j}(u)\begin{pmatrix} \frac{1}{2}\\ j \end{pmatrix}a_1^{j}(u)\nonumber \\&\quad =\cos ^{2\tau }(u)\sum _{k=0}^{2\tau }\begin{pmatrix} \frac{1}{2}\\ k \end{pmatrix}\left( -\frac{15}{16}\right) ^k+\sum _{j=1}^{\tau }\sum _{k=0}^{2\tau -2j}\begin{pmatrix} \frac{1}{2}-j\\ k \end{pmatrix}b_1^k(u)\cos ^{2\tau -k-2j}(u)\begin{pmatrix} \frac{1}{2}\\ j \end{pmatrix}a_1^{j}(u)\nonumber \\&\quad \ge \cos ^{2\tau }(u)\sqrt{1-\frac{15}{16}}+\sum _{j=1}^{\tau }\left[ \sum _{k=0}^{2\tau -2j}\begin{pmatrix} \frac{1}{2}-j\\ k \end{pmatrix}b_1^k(u)\right] \begin{pmatrix} \frac{1}{2}\\ j \end{pmatrix}a_1^{j}(u)\nonumber \\&\quad =(1+b_1(u))^{\frac{1}{2}}\sum _{j=1}^{\infty }\begin{pmatrix} \frac{1}{2}\\ j \end{pmatrix}\left( \frac{a_1(u)}{1+b_1(u)}\right) ^{j}\nonumber \\&\quad =(1+b_1(u))^{\frac{1}{2}}\left( -1+\sqrt{1+\frac{a_1(u)}{1+b_1(u)}}\right) \nonumber \\&\quad =\frac{a_1(u)}{(1+b_1(u))^{\frac{1}{2}}}\frac{1}{1+\sqrt{1-\frac{-a_1(u)}{1+b_1(u)}}}. \end{aligned}$$

Since the function \(\frac{a_1(u)}{(1+b_1(u))^{\frac{1}{2}}}\) is negative and increasing in \((0,\frac{\pi }{2})\) (because its derivative is equal to \( \frac{ \left( 45 \cos ^{2}{\left( u \right) } - 64 \cos {\left( u \right) } + 45\right) \sin {\left( u \right) }}{16 \left( 16 -15 \cos {\left( u \right) }\right) ^{\frac{3}{2}}} \), which is positive for all \(u\in (0,\frac{\pi }{2})\)), and the function \(\frac{-a_1(u)}{1+b_1(u)}\) is positive and decreasing in \((0,\frac{\pi }{2})\) (since its derivative \(- \frac{\left( 15 \cos ^{2}{\left( u \right) } - 32 \cos {\left( u \right) } + 45\right) \sin {\left( u \right) }}{2 \left( 15 \cos {\left( u \right) } - 16\right) ^{2}}\) is negative for \(u\in (0,\frac{\pi }{2})\)), we have that the last expression in the chain of inequalities mentioned above is increasing. Then, it reaches its minimum in \(\left[ 0,\frac{\pi }{2}\right] \) at \(u=0\). Hence \(c_{1,2\tau }(u)\ge c_{1,2\tau }(0)= - \frac{1}{4}\) for every \(u\in (0,\frac{\pi }{2})\). Therefore

$$\begin{aligned} \int _{0}^{\pi }c_{1,2\tau }(u)\mathrm{d}u=2\int _{0}^{\frac{\pi }{2}}c_{1,2\tau }(u)\mathrm{d}u\ge - \frac{\pi }{4}. \end{aligned}$$

\(\square \)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Beltritti, G. Periodic solutions of a generalized Sitnikov problem. Celest Mech Dyn Astr 133, 6 (2021).

Download citation


  • Generalized Sitnikov problem
  • Even periodic solutions
  • Global families
  • Continuation from circular Sitnikov problem