Namespaces
Variants
Actions

Difference between revisions of "Arc (projective geometry)"

From Encyclopedia of Mathematics
Jump to: navigation, search
(→‎References: title rank)
(finish TeX)
 
(2 intermediate revisions by one other user not shown)
Line 1: Line 1:
 +
<!--This article has been texified automatically. Since there was no Nroff source code for this article,
 +
the semi-automatic procedure described at https://encyclopediaofmath.org/wiki/User:Maximilian_Janisch/latexlist
 +
was used.
 +
If the TeX and formula formatting is correct and if all png images have been replaced by TeX code, please remove this message and the {{TEX|semi-auto}} category.
 +
 +
Out of 96 formulas, 95 were replaced by TEX code.-->
 +
 +
{{TEX|done}}
 
==$k$-arcs in projective planes==
 
==$k$-arcs in projective planes==
A <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a1202504.png" />-arc in the Desarguesian [[Projective plane|projective plane]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a1202505.png" /> over the [[Galois field|Galois field]] of order <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a1202506.png" /> is a set of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a1202507.png" /> points, no three of which are collinear. It is immediate that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a1202508.png" />, but if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a1202509.png" /> is odd, then <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025010.png" />. The classical example of a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025011.png" />-arc is a conic, that is, a set of points projectively equivalent to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025012.png" />. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025013.png" /> is even, then all the tangents to the conic pass through a common point, called the nucleus; hence the set of points of a conic together with the nucleus is a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025014.png" />-arc. A <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025015.png" />-arc in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025016.png" /> is called an oval and a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025017.png" />-arc in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025018.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025019.png" /> even, is called a hyperoval (cf. also [[Oval|Oval]]).
+
A $k$-arc in the Desarguesian [[Projective plane|projective plane]] $\operatorname{PG} ( 2 , q )$ over the [[Galois field|Galois field]] of order $q$ is a set of $k$ points, no three of which are collinear. It is immediate that $k \leq q + 2$, but if $q$ is odd, then $k \leq q + 1$. The classical example of a $( q + 1 )$-arc is a conic, that is, a set of points projectively equivalent to $\{ ( 1 , t , t ^ { 2 } ) : t \in \operatorname {GF} ( q ) \} \cup \{ ( 0,0,1 ) \}$. If $q$ is even, then all the tangents to the conic pass through a common point, called the nucleus; hence the set of points of a conic together with the nucleus is a $( q + 2 )$-arc. A $( q + 1 )$-arc in $\operatorname{PG} ( 2 , q )$ is called an oval and a $( q + 2 )$-arc in $\operatorname{PG} ( 2 , q )$, $q$ even, is called a hyperoval (cf. also [[Oval|Oval]]).
  
In his celebrated 1955 theorem, B. Segre proved that in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025020.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025021.png" /> odd, every <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025022.png" />-arc is a conic. This important result, linking combinatorial and algebraic properties of sets of points, was of great importance in the early development of the field of finite geometry, and many results in the same spirit have been proved.
+
In his celebrated 1955 theorem, B. Segre proved that in $\operatorname{PG} ( 2 , q )$, $q$ odd, every $( q + 1 )$-arc is a conic. This important result, linking combinatorial and algebraic properties of sets of points, was of great importance in the early development of the field of finite geometry, and many results in the same spirit have been proved.
  
The situation when <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025023.png" /> is even is quite different. Apart from the  "classical"  examples provided by a conic together with its nucleus mentioned above, there are currently (1998) seven infinite families of hyperovals known, and several examples which do not at present fit into any known infinite family. The classification of hyperovals is known only for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025024.png" />; the case <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025025.png" /> still relies on a computer search.
+
The situation when $q$ is even is quite different. Apart from the  "classical"  examples provided by a conic together with its nucleus mentioned above, there are currently (1998) seven infinite families of hyperovals known, and several examples which do not at present fit into any known infinite family. The classification of hyperovals is known only for $q \leq 32$; the case $q = 32$ still relies on a computer search.
  
 
==$k$-arcs in projective space==
 
==$k$-arcs in projective space==
A <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025029.png" />-arc in the <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025030.png" />-dimensional [[Projective space|projective space]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025031.png" /> is a set of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025032.png" /> points with <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025033.png" /> and at most <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025034.png" /> in each hyperplane. (This definition coincides with the definition of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025035.png" />-arc in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025036.png" /> given above.) Further, a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025037.png" />-arc of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025038.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025039.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025040.png" />, exists if and only if a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025041.png" />-arc of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025042.png" /> exists. A linear maximum distance separable code is a linear code of length <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025043.png" />, dimension <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025044.png" /> and minimum distance <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025045.png" /> (cf. also [[Coding and decoding|Coding and decoding]]). It is well-known that for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025046.png" /> these notions are equivalent; as each can be viewed as a set of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025047.png" /> vectors in an <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025048.png" />-dimensional [[Vector space|vector space]] over <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025049.png" /> with each <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025050.png" /> vectors being linearly independent.
+
A $k$-arc in the $n$-dimensional [[Projective space|projective space]] $\operatorname{PG} ( n , q )$ is a set of $k$ points with $k \geq n + 1$ and at most $n$ in each hyperplane. (This definition coincides with the definition of $k$-arc in $\operatorname{PG} ( 2 , q )$ given above.) Further, a $k$-arc of $\operatorname{PG} ( n , q )$, $n \geq 2$ and $k \geq n + 4$, exists if and only if a $k$-arc of $\operatorname{PG} ( k - n - 2 , q )$ exists. A linear maximum distance separable code is a linear code of length $k$, dimension $n + 1$ and minimum distance $d = k - n + 2$ (cf. also [[Coding and decoding|Coding and decoding]]). It is well-known that for $k \geq n + 1$ these notions are equivalent; as each can be viewed as a set of $k$ vectors in an $( n + 1 )$-dimensional [[Vector space|vector space]] over $\operatorname{GF} ( q ),$ with each $n + 1$ vectors being linearly independent.
  
The classical example of a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025051.png" />-arc in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025052.png" /> is a normal rational curve, that is, a set of points projectively equivalent to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025053.png" />. The only known (1998) non-classical examples are examples in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025054.png" />, for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025055.png" /> even, constructed by L.R.A. Casse and D.G. Glynn, and a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025056.png" />-arc in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025057.png" /> constructed by Glynn. The main open (1998) problem in the area is the resolution of the so-called main conjecture for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025059.png" />-arcs and maximum distance separable codes, which is that if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025060.png" />, then the size of a largest <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025061.png" />-arc in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025062.png" /> is <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025063.png" /> if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025064.png" /> is even and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025065.png" /> or <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025066.png" />, and is <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025067.png" /> in all other cases. The main conjecture has been settled for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025068
+
The classical example of a $( q + 1 )$-arc in $\operatorname{PG} ( n , q )$ is a normal rational curve, that is, a set of points projectively equivalent to $\{ ( 1 , t , t ^ { 2 } , \dots , t ^ { n } ) : t \in \operatorname{GF} ( q ) \} \cup \{ ( 0 , \dots , 0,1 ) \}$. The only known (1998) non-classical examples are examples in $\operatorname{PG} ( 3 , q )$, for $q$ even, constructed by L.R.A. Casse and D.G. Glynn, and a $10$-arc in $\operatorname{PG} ( 4,9 )$ constructed by Glynn. The main open (1998) problem in the area is the resolution of the so-called main conjecture for $k$-arcs and maximum distance separable codes, which is that if $q > n + 1$, then the size of a largest $k$-arc in $\operatorname{PG} ( n , q )$ is $q + 2$ if $q$ is even and $n = 2$ or $n = q - 2$, and is $q + 1$ in all other cases. The main conjecture has been settled for $n=2,3,4$ and in a number of further cases. It is also of great interest to characterize the largest-size $k$-arcs, and to determine the size of the second-largest complete $k$-arcs in $\operatorname{PG} ( n , q )$, where a $k$-arc of $\operatorname{PG} ( n , q )$ is complete if it is contained in no $( k + 1 )$-arc of $\operatorname{PG} ( n , q )$.
.png" /> and in a number of further cases. It is also of great interest to characterize the largest-size <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025069.png" />-arcs, and to determine the size of the second-largest complete <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025070.png" />-arcs in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025071.png" />, where a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025072.png" />-arc of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025073.png" /> is complete if it is contained in no <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025075.png" />-arc of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025076.png" />.
 
  
 
==$(k,n)$-arcs in $\mathrm{PG}(2,q)$==
 
==$(k,n)$-arcs in $\mathrm{PG}(2,q)$==
For <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025079.png" />, a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025081.png" />-arc (or arc of degree <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025083.png" />) in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025084.png" /> is a set <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025085.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025086.png" /> points such that each line meets <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025087.png" /> in at most <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025088.png" /> points and there is a line meeting <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025089.png" /> in exactly <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025090.png" /> points. It is immediate that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025091.png" /> with equality if and only if each line meets <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025092.png" /> in 0 or <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025093.png" /> points. Equality also implies that either <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025094.png" /> or <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025095.png" /> divides <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025096.png" />. A <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025097.png" />-arc with <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a12025098.png" /> is called a maximal arc of degree <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250100.png" />, and is non-trivial if <img align="absmiddle" bor
+
For $n \geq 2$, a $( k , n )$-arc (or arc of degree $n$) in $\operatorname{PG} ( 2 , q )$ is a set $\mathcal{K}$ of $k$ points such that each line meets $\mathcal{K}$ in at most $n$ points and there is a line meeting $\mathcal{K}$ in exactly $n$ points. It is immediate that $k \leq ( n - 1 ) q + n,$ with equality if and only if each line meets $\mathcal{K}$ in 0 or $n$ points. Equality also implies that either $n = q + 1$ or $n$ divides $q$. A $( k , n )$-arc with $k = ( n - 1 ) q + n$ is called a maximal arc of degree $n$, and is non-trivial if $2 \leq n \leq q - 1$. If $q$ is even, there are examples of non-trivial maximal arcs of degree $n$ for every $n$ dividing $q$ (due to R.H.F. Denniston and J.A. Thas). On the other hand, S. Ball, A. Blokhuis and F. Mazzocca have shown that non-trivial maximal arcs in $\operatorname {PG} ( 2 , q ),$ where $q$ is odd, do not exist. (The proof appears in [[#References|[a2]]].)
der="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250101.png" />. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250102.png" /> is even, there are examples of non-trivial maximal arcs of degree <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250103.png" /> for every <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250104.png" /> dividing <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250105.png" /> (due to R.H.F. Denniston and J.A. Thas). On the other hand, S. Ball, A. Blokhuis and F. Mazzocca have shown that non-trivial maximal arcs in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250106.png" /> where <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250107.png" /> is odd, do not exist. (The proof appears in [[#References|[a2]]].)
 
  
 
See [[#References|[a3]]] for a survey on each topic mentioned above; for a comprehensive account including more details, results and the references, see [[#References|[a2]]], Chaps. 8, 9.
 
See [[#References|[a3]]] for a survey on each topic mentioned above; for a comprehensive account including more details, results and the references, see [[#References|[a2]]], Chaps. 8, 9.
  
The definitions of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250108.png" />-arc, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/a/a120/a120250/a120250109.png" />-arc, oval, hyperoval and maximal arc in a non–Desarguesian projective plane are the combinatorial definitions given above, but in this case there are relatively few examples and the theory is not so well-developed.
+
The definitions of $k$-arc, $( k , n )$-arc, oval, hyperoval and maximal arc in a non–Desarguesian projective plane are the combinatorial definitions given above, but in this case there are relatively few examples and the theory is not so well-developed.
  
 
==References==
 
==References==
<table><TR><TD valign="top">[a1]</TD> <TD valign="top">  J.W.P. Hirschfeld,  J.A. Thas,  "General Galois geometries" , Oxford Univ. Press  (1991)</TD></TR><TR><TD valign="top">[a2]</TD> <TD valign="top">  J.W.P. Hirschfeld,  "Projective geometries over finite fields" , Oxford Univ. Press  (1998)  (Edition: Second)</TD></TR><TR><TD valign="top">[a3]</TD> <TD valign="top">  J.A. Thas,  "Projective geometry over a finite field"  F. Buekenhout (ed.) , ''Handbook of Incidence Geometry, Buildings and Foundations'' , Elsevier  (1995)  pp. Chap. 7; 295–348</TD></TR></table>
+
<table><tr><td valign="top">[a1]</td> <td valign="top">  J.W.P. Hirschfeld,  J.A. Thas,  "General Galois geometries" , Oxford Univ. Press  (1991)</td></tr><tr><td valign="top">[a2]</td> <td valign="top">  J.W.P. Hirschfeld,  "Projective geometries over finite fields" , Oxford Univ. Press  (1998)  (Edition: Second)</td></tr><tr><td valign="top">[a3]</td> <td valign="top">  J.A. Thas,  "Projective geometry over a finite field"  F. Buekenhout (ed.) , ''Handbook of Incidence Geometry, Buildings and Foundations'' , Elsevier  (1995)  pp. Chap. 7; 295–348</td></tr></table>

Latest revision as of 18:09, 5 August 2020

$k$-arcs in projective planes

A $k$-arc in the Desarguesian projective plane $\operatorname{PG} ( 2 , q )$ over the Galois field of order $q$ is a set of $k$ points, no three of which are collinear. It is immediate that $k \leq q + 2$, but if $q$ is odd, then $k \leq q + 1$. The classical example of a $( q + 1 )$-arc is a conic, that is, a set of points projectively equivalent to $\{ ( 1 , t , t ^ { 2 } ) : t \in \operatorname {GF} ( q ) \} \cup \{ ( 0,0,1 ) \}$. If $q$ is even, then all the tangents to the conic pass through a common point, called the nucleus; hence the set of points of a conic together with the nucleus is a $( q + 2 )$-arc. A $( q + 1 )$-arc in $\operatorname{PG} ( 2 , q )$ is called an oval and a $( q + 2 )$-arc in $\operatorname{PG} ( 2 , q )$, $q$ even, is called a hyperoval (cf. also Oval).

In his celebrated 1955 theorem, B. Segre proved that in $\operatorname{PG} ( 2 , q )$, $q$ odd, every $( q + 1 )$-arc is a conic. This important result, linking combinatorial and algebraic properties of sets of points, was of great importance in the early development of the field of finite geometry, and many results in the same spirit have been proved.

The situation when $q$ is even is quite different. Apart from the "classical" examples provided by a conic together with its nucleus mentioned above, there are currently (1998) seven infinite families of hyperovals known, and several examples which do not at present fit into any known infinite family. The classification of hyperovals is known only for $q \leq 32$; the case $q = 32$ still relies on a computer search.

$k$-arcs in projective space

A $k$-arc in the $n$-dimensional projective space $\operatorname{PG} ( n , q )$ is a set of $k$ points with $k \geq n + 1$ and at most $n$ in each hyperplane. (This definition coincides with the definition of $k$-arc in $\operatorname{PG} ( 2 , q )$ given above.) Further, a $k$-arc of $\operatorname{PG} ( n , q )$, $n \geq 2$ and $k \geq n + 4$, exists if and only if a $k$-arc of $\operatorname{PG} ( k - n - 2 , q )$ exists. A linear maximum distance separable code is a linear code of length $k$, dimension $n + 1$ and minimum distance $d = k - n + 2$ (cf. also Coding and decoding). It is well-known that for $k \geq n + 1$ these notions are equivalent; as each can be viewed as a set of $k$ vectors in an $( n + 1 )$-dimensional vector space over $\operatorname{GF} ( q ),$ with each $n + 1$ vectors being linearly independent.

The classical example of a $( q + 1 )$-arc in $\operatorname{PG} ( n , q )$ is a normal rational curve, that is, a set of points projectively equivalent to $\{ ( 1 , t , t ^ { 2 } , \dots , t ^ { n } ) : t \in \operatorname{GF} ( q ) \} \cup \{ ( 0 , \dots , 0,1 ) \}$. The only known (1998) non-classical examples are examples in $\operatorname{PG} ( 3 , q )$, for $q$ even, constructed by L.R.A. Casse and D.G. Glynn, and a $10$-arc in $\operatorname{PG} ( 4,9 )$ constructed by Glynn. The main open (1998) problem in the area is the resolution of the so-called main conjecture for $k$-arcs and maximum distance separable codes, which is that if $q > n + 1$, then the size of a largest $k$-arc in $\operatorname{PG} ( n , q )$ is $q + 2$ if $q$ is even and $n = 2$ or $n = q - 2$, and is $q + 1$ in all other cases. The main conjecture has been settled for $n=2,3,4$ and in a number of further cases. It is also of great interest to characterize the largest-size $k$-arcs, and to determine the size of the second-largest complete $k$-arcs in $\operatorname{PG} ( n , q )$, where a $k$-arc of $\operatorname{PG} ( n , q )$ is complete if it is contained in no $( k + 1 )$-arc of $\operatorname{PG} ( n , q )$.

$(k,n)$-arcs in $\mathrm{PG}(2,q)$

For $n \geq 2$, a $( k , n )$-arc (or arc of degree $n$) in $\operatorname{PG} ( 2 , q )$ is a set $\mathcal{K}$ of $k$ points such that each line meets $\mathcal{K}$ in at most $n$ points and there is a line meeting $\mathcal{K}$ in exactly $n$ points. It is immediate that $k \leq ( n - 1 ) q + n,$ with equality if and only if each line meets $\mathcal{K}$ in 0 or $n$ points. Equality also implies that either $n = q + 1$ or $n$ divides $q$. A $( k , n )$-arc with $k = ( n - 1 ) q + n$ is called a maximal arc of degree $n$, and is non-trivial if $2 \leq n \leq q - 1$. If $q$ is even, there are examples of non-trivial maximal arcs of degree $n$ for every $n$ dividing $q$ (due to R.H.F. Denniston and J.A. Thas). On the other hand, S. Ball, A. Blokhuis and F. Mazzocca have shown that non-trivial maximal arcs in $\operatorname {PG} ( 2 , q ),$ where $q$ is odd, do not exist. (The proof appears in [a2].)

See [a3] for a survey on each topic mentioned above; for a comprehensive account including more details, results and the references, see [a2], Chaps. 8, 9.

The definitions of $k$-arc, $( k , n )$-arc, oval, hyperoval and maximal arc in a non–Desarguesian projective plane are the combinatorial definitions given above, but in this case there are relatively few examples and the theory is not so well-developed.

References

[a1] J.W.P. Hirschfeld, J.A. Thas, "General Galois geometries" , Oxford Univ. Press (1991)
[a2] J.W.P. Hirschfeld, "Projective geometries over finite fields" , Oxford Univ. Press (1998) (Edition: Second)
[a3] J.A. Thas, "Projective geometry over a finite field" F. Buekenhout (ed.) , Handbook of Incidence Geometry, Buildings and Foundations , Elsevier (1995) pp. Chap. 7; 295–348
How to Cite This Entry:
Arc (projective geometry). Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Arc_(projective_geometry)&oldid=25358