Finite projective planes and the Delsarte LP-bound (Q1748012)

All projective planes of order up to 10 have been classified. Here, some of these results are reproved. Specifically, it is shown that there is no projective plane of order 6 and that there is a unique projective plane of order 7. The existence of a projective plane of order \(n\) is equivalent to the existence of \(n-1\) mutually orthogonal Latin squares (MOLS) of side \(n\). This connects the current study to the nonexistence of 2 MOLS of side 6. That result was originally obtained by \textit{G. Tarry} [Ass. Franç. Paris 29, 170--203 (1900; JFM 32.0219.04)] more than a century ago. About ten alternative proofs have later been obtained, including those by \textit{D. R. Stinson} [J. Comb. Theory, Ser. A 36, 373--376 (1984; Zbl 0538.05012)], \textit{S. T. Dougherty} [Des. Codes Cryptography 4, No. 2, 123--128 (1994; Zbl 0796.05012)], \textit{G. Appa} et al. [Oper. Res. Lett. 32, No. 4, 336--344 (2004; Zbl 1052.05018)], and \textit{A. P. Burger} et al. [Discrete Math. 311, No. 13, 1223--1228 (2011; Zbl 1229.05050)]. Appa et al. [loc. cit.] use linear programming (LP), which is also the approach of the current paper. The current study would hence have benefited from a comparison with those old results. One well-known difference between a search for a small number of MOLS of side \(n\) and a search for \(n-1\) MOLS of side \(n\) (that is, a projective plane) is that in the latter case one may actually start from a substructure coming from a Latin square of side \(n-1\) rather than \(n\); this observation is utilized by the authors. In some cases, a partial structure has to be extended somewhat, through a very small search tree, before LP gives the desired nonexistence result; also this is in line with the results of Appa et al. [loc. cit.].