Abstract
We propose a novel optimization algorithm for differentiable functions utilizing geodesics and contours under conformal mapping. The algorithm can locate multiple optima by first following a geodesic curve to a local optimum then traveling to the next search area by following a contour curve. Alongside we implement a jumping mechanism which we call shadow casting to help geodesics jump to locations closer to the global optimum. To improve the efficiency, local search methods such as the Newton–Raphson algorithm are also employed. For functions with many optima or when the global optimum is very close to a local one, numerical analyses have shown that the resulting algorithm, SGEO-QN, can outperform recent derivative-free DIRECT variants in number of function/gradient evaluations. The results also indicate that under certain conditions, number of function/gradient evaluations for SGEO-QN scales nearly linearly with increasing dimensionality. Lastly, SGEO-QN appears to be less affected by rotational transforms of the objective functions than the variants of DIRECT compared.
Similar content being viewed by others
Notes
We use a shorthand for partial derivatives, \(\partial _i \equiv \frac{\partial }{\partial x^i}\).
Here, \(\mathbf {v}\) and \(\nabla f\) represent the tangent and gradient vectors in Euclidean space.
Note that \( \sum _t f(\mathbf {x}_t) \ne 1\) and the vector \(\sum _t \widehat{f}(\mathbf {x}_t) (\mathbf {x}_t - \mathbf {A})\) will be approximately parallel to \((\mathbf {x}_t - \mathbf {A})\) in Eq. (6).
Note that \(\mathbf {x}^*_{\gamma } \in L\).
This is similar to the monotonic basin hopping algorithm [27] but we use a random walk instead of uniform sampling in a D-sphere.
We implement quasi-Newton (QN) local searches on the geodesics \(\gamma \) and the line L. Hence we call the resulting algorithm SGEO-QN.
A recent publication tackling the problem of comparing deterministic and stochastic optimization methods is discussed in [34].
The only quantity that depends on the coordinate is \(t_C\) in Eq. (9), where a different choice of basis would result in different values in the components.
The composition functions have the form \(f(x) = \sum _i \omega _i g_i(x)\), where i denotes the i-th basic function \(g_i(x)\) and its weight \(\omega _i \in [0,1]\), such that \(\sum _i \omega _i = 1\).
References
Diener, I.: Handbook of Global Optimization, Vol. 2 of the Series Nonconvex Optimization and its Applications, pp. 649–668 (1995)
Floudas, C.A., Pardalos, P. M.: Encyclopedia of Optimization, Globally Convergent Homotopy Methods, pp. 1272–1277
Chow, S.N., Mallet-Paret, J., Yorke, J.A.: Finding zeros of maps: homotopy methods that are constructive with probability one. Math. Comput. 32, 887899 (1978)
Watson, L.T.: Globally convergent homotopy algorithms for nonlinear systems of equations. Nonlinear Dyn. 1, 143–191 (1990)
Dunlavy, D.M., O’Leary, D.P.: Homotopy Optimization Methods for Global Optimization, Sandia National Laboratories, Report SAND 2005–7495 (2005)
Botsaris, C.A.: Constrained optimization along geodesics. J. Math. Anal. Appl. 79(2), 295–306 (1981)
Smith, S.T.: Geometric Optimization Methods for Adaptive Filtering. Harvard University, Cambridge (1993)
Smith, S.T.: Optimization techniques on Riemannian manifolds. Fields Inst. Commun. 3(3), 113–135 (1994)
Rapcsák, T.: Convex Programming on Riemannian Manifolds, Lecture Notes in Control and Information Sciences, vol. 84, pp. 733–740. Springer, Berlin (1986)
Rapcsák, T.: Geodesic convexity in nonlinear optimization. J. Optim. Theory Appl. 69(1), 169–183 (1991)
Rapcsák, T.: Geodesic convexity on \(\mathbb{R}^n_+\). Optimization 37(4), 341–355 (1996)
Rapcsák, T.: Local convexity on smooth manifolds. J. Optim. Theory Appl. 127(1), 165–176 (2005)
Perttunen, C.D., Jones, D.R., Stuckman, B.E.: Lipschitzian optimization without the Lipschitz constant. J. Optim. Theory Appl. 79(1), 157181 (1993)
Liuzzi, G., Lucidi, S., Piccialli, V.: A DIRECT-based approach exploiting local minimizations for the solution of large-scale global optimization problems. Comput. Optim. Appl. 45(2), 353–375 (2010)
Liuzzi, G., Lucidi, S., Piccialli, V.: Exploiting derivative-free local searches in DIRECT-type algorithms for global optimization. Comput. Optim. Appl. (2015). doi:10.1007/s10589-015-9741-9
Kvasov, D.E., Sergeyev, Y.D.: A univariate global search working with a set of Lipschitz constants for the first derivative. Optim. Lett. 3(2), 303–318 (2009)
Kvasov, D.E., Sergeyev, Y.D.: Lipschitz gradients for global optimization in a one-point-based partitioning scheme. J. Comput. Appl. Math. 236(16), 4042–4054 (2012)
Butz, A.R.: Space-filling curves and mathematical programming. Inf. Control 12(4), 313–330 (1968)
Lera, D., Sergeyev, Y.D.: Deterministic global optimization using space-filling curves and multiple estimates of Lipschitz and Hölder constants. Commun. Nonlinear Sci. Numer. Simul. 23(1–3), 328–342 (2015)
Sergeyev, Y.D., Strongin, R.G., Lera, D.: Introduction to Global Optimization Exploiting Space-Filling Curves. Springer, Berlin (2013). doi:10.1007/978-1-4614-8042-6
Schoen, F.: Two-Phase Methods for Global Optimization, Handbook of Global Optimization, Vol. 62 of the Series Nonconvex Optimization and its Applications, pp. 151–177 (2002)
Sergeyev, Y., Kvasov, D.E.: Global search based on efficient diagonal partitions and a set of Lipschitz constants. SIAM J. Optim. 16(3), 910937 (2006). doi:10.1137/040621132
Boothby, M.W.: An Introduction to Differentiable Manifolds and Riemannian Geometry. Springer, Berlin (2003)
Lee, M.J.: Introduction to Topological Manifolds. Springer, Berlin (2010)
Foster, J.: A Short Course in General Relativity. Springer, Berlin (2006)
Hawking, S., Ellis, G.F.R.: The Large Scale Structure of Space-Time. Cambridge University Press, Cambridge (1973)
Wales, D.J., Doye, J.P.K.: Global optimization by basin-hopping and the lowest energy structures of Lennard-Jones Clusters containing up to 110 atoms. Phys. Chem. A 101, 5111 (1997)
Gaviano, M., Kvasov, D.E., Lera, D., Sergeyev, Y.D.: Algorithm 829: software for generation of classes of test functions with known local and global minima for global optimization. ACM Trans. Math. Softw. 29(4), 469–480 (2003)
Addis, B., Locatelli, M.: A new class of test functions for global optimization. J. Glob. Optim. 38(3), 479–501 (2007)
Zilinskas, A.: A class of test functions for global optimization. J. Glob. Optim. 5(2), 195–199 (1994)
Schoen, F.: A wide class of test functions for global optimization. J. Glob. Optim. 3, 133–138 (1993)
Hedar, A.R.: Test Problems for Unconstrained Optimization, http://www-optima.amp.i.kyoto-u.ac.jp/member/student/hedar/Hedar_files/TestGO_files/Page364.htm
Liang, J.J., Qu, B.Y., Suganthan, P.N., Hernndaz-Daz, A.G.: Problem Definitions and Evaluation Criteria for the CEC 2013 Special Session and Competition on Real-parameter Optimization, Technical Report 201212. Computational Intelligence Laboratory, Zhengzhou University, Zhengzhou China (2013)
Sergeyev, Y., Kvasov, D.E., Mukhametzhanov, M.S.: Operational zones for comparing metaheuristic and deterministic one-dimensional global optimization algorithms. Math. Comput. Simul. (2016). doi:10.1016/j.matcom.2016.05.006
Jamil, M., Yang, X.S.: A literature survey of benchmark functions for global optimisation problems. Int. J. Math. Model. Numer. Optim. 4(2), 150–194 (2013)
Acknowledgments
We thank the anonymous reviewers for extremely useful feedbacks in assisting the writing of this manuscript.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflicts of interest
The authors declare that they have no conflict of interest.
Funding
This research is supported by the Discovery Grants and Discovery Accelerator Supplement from Natural Sciences and Engineering Research Council of Canada (NSERC).
Rights and permissions
About this article
Cite this article
Fok, R., An, A. & Wang, X. Geodesic and contour optimization using conformal mapping. J Glob Optim 69, 23–44 (2017). https://doi.org/10.1007/s10898-016-0467-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10898-016-0467-8