Classifying (almost)-Belyi maps with five exceptional points (Q1629131)

The authors' main aim to write this paper is to solve second order linear differential equations with rational coefficients having \((d=5)\) exceptional points (true singularities) in terms of hypergeometric functions and thus prepare a database (table) `up to Möbius-equivalence' [see, \textit{M. van Hoeij} and \textit{V. J. Kunwar}, Classifying (almost)-Belyi maps with Five Exceptional Points. Data, algorithms, and examples, \url{http://www.math.fsu.edu/~hoeij/FiveSing.}] of all rational functions \(f \in \mathbb{C}(x)\) `with five \((\infty, 2, 3)\)-exceptional points, and likewise for \((\infty, 2, 4)\) and \((\infty, 2, 6)\)' and to prove that this table is complete and to find for a given \(k \subseteq \mathbb{C}\) and \(\{q_1,\dots,q_5\} \subset \mathbb{P}^1\) every \(F \in k(x)\) such that \(E(F)=\{q_1,\dots,q_5\}\) in which \(E\) denotes the set of exceptional points of \(f\). This work is in continuation of the previous works of the first author with his co-workers (see, [\textit{M. van Hoeij} and \textit{R. Vidunas}, J. Algebra 441, 609--659 (2015; Zbl 1332.33036)], [\textit{V. J. Kunwar}, Hypergeometric solutions of linear differential equations with rational function coefficients (Ph.D. thesis) (2014), \url{http://www.math.fsu.edu/~vkunwar}; \textit{V. J. Kunwar} and \textit{M. van Hoeij}, in: Proceedings of the 38th international symposium on symbolic and algebraic computation, ISSAC 2013, Boston, MA, USA, June 26--29, 2013. New York, NY: Association for Computing Machinery (ACM). 235--242 (2013; Zbl 1360.34007)], [\textit{R. Vidūnas}, J. Comput. Appl. Math. 178, No. 1--2, 473--487 (2005; Zbl 1076.33002); \textit{R. Vidūnas} and \textit{G. Filipuk}, Funkc. Ekvacioj, Ser. Int. 56, No. 2, 271--321 (2013; Zbl 1279.33030); \textit{R. Vidūnas} and \textit{A. V. Kitaev}, Math. Comput. 78, No. 268, 2371--2395 (2009; Zbl 1226.14043)] regarding solutions of differential equations in terms of hypergeometric functions yet it differs from these in certain aspects like, a powerful algorithm (Algorithm 4.4) is developed here for evaluating dessins to prove the completeness of the table. While the study [van Hoeij and R. Vidunas, loc. cit.] dealt with methods specific to the \(d=4\) case without using dessins and while the tables developed in the work [Vidūnas and Filipuk, loc. cit.] were `small enough for manual enumeration'. Further, the authors here `compute almost-Belyi maps and braid orbits of almost-dessins' to show the completness of their table in contrast to the previous works of van Hoeij and Vidūnas [loc. cit.] and Vidūnas and Filipuk [loc. cit.] which dealt only with Belyi maps. The most important algorithm of the paper is the Algorithm 4.4 which computes dessins (conjugacy class of 3-constellations) to establish that the Belyi table developed by the authors is complete. This algorithm introduced by the authors here is better than the existing algorithm of \textit{J. Sijsling} [Belyi maps: explicit determination of the permutation triples of given ramification type. (2010), \url{sites.google.com/site/sijsling/programs}] which is `not efficient enough for higher degrees'. Prior to arriving at this key algorithm of the paper, the authors give three more precursor algorithms and a proposition in this section, in which the Algorithm 4.3 is used for verifying that `the 3-constellations \([g_0, g_1, g_{\infty}]\) and \([h_0, h_1, h_{\infty}]\) are conjugated'. By considering planar 4-constellations \([g_0, g_1, g_t , g_{\infty}]\) for a 2-cycle \(g_t \) the authors compute almost-dessins (which correspond to \(\mathrm{Bely}\mathrm{i}^{(1)}\) maps up to Möbius-equivalence) in the fifth section of the paper, where they compute braid orbits of almost-dessins by combining the algorithm given in [\textit{K. Magaard} et al., Exp. Math. 12, No. 4, 385--393 (2003; Zbl 1068.12002)] for computing braid orbits of \(k\)-constellations \([g_1,\dots,g_k]\) with their Algorithm 4.3. The \(\mathrm{Bely}\mathrm{i}^{(2)}\) maps are discussed in the sixth section of the paper and the seventh section is devoted to a short discussion on five point invariants. Finally linear differential equations with five true singularities (at least one of which is logarithmic) are solved in terms of the Guass's hypergeometric function \({}_2{F_1}\) in the eighth section of the paper and the method is also illustrated with the help of an example. The reviewer strongly feels that the complete table of Belyi maps with five exceptional points developed in this paper will be useful in other fields of applications of these maps like combinatorics, algebra, geometry, etc.