For various values of n and m we investigate homomorphisms from Out(F_n) to Out(F_m) and from Out(F_n) to GL_m(K), i.e. the free and linear representations of Out(F_n) respectively. By means of a series of arguments revolving around the representation theory of finite symmetric subgroups of Out(F_n) we prove that each homomorphism from Out(F_n) to GL_m(K) factors through the natural map p_n from Out(F_n) to GL(H_1(F_n,Z)) = GL_n(Z) whenever n=3, m < 7 and char(K) is not an element of {2,3}, and whenever n>5, m< n(n+1)/2 and char(K) is not an element of {2,3,...,n+1}. We also construct a new infinite family of linear representations of Out(F_n) (where n > 2), which do not factor through p_n. When n is odd these have the smallest dimension among all known representations of Out(F_n) with this property. Using the above results we establish that the image of every homomorphism from Out(F_n) to Out(F_m) is finite whenever n=3 and n < m < 6, and of cardinality at most 2 whenever n > 5 and n < m < n(n-1)/2. We further show that the image is finite when n(n-1)/2 -1 < m < n(n+1)/2. We also consider the structure of normal finite index subgroups of Out(F_n). If N is such then we prove that if the derived subgroup of the intersection of N with the Torelli subgroup T_n < Out(F_n) contains some term of the lower central series of T_n then the abelianisation of N is finite.

Let F be a free group of finite rank. Given words u, v ∈ F, J.H.C. Whitehead solved the decision problem of finding an automorphism φ ∈ Aut(F), carrying u to v. He used topological methods to produce an algorithm. Higgins and Lyndon gave a very concise proof of the problem based on the works of Rapaport. We provide a detailed account of Higgins and Lyndon’s proof of the peak reduction lemma and the restricted version of Whitehead’s theorem, for cyclic words as well as for sets of cyclic words, with a full explanation of each step. Then, we give an inductive proof of Whitehead’s minimization theorem and describe Whitehead’s decision algorithm. Noticing that Higgins and Lyndon’s work is limited to the cyclic words, we extend their proofs to ordinary words and sets of ordinary words. In the last chapter, we mention an example given by Whitehead to show that the decision problem for finitely generated subgroups is more difficult and outline an approach due to Gersten to overcome this difficulty. We also give an extensive literature survey of Whitehead’s algorithm / Thesis / Master of Science (MSc)

Un automorphisme (extérieur) $phi $ d'un groupe libre $F_n$ de rang fini $ngeq 2$ est dit géométrique s'il est induit par un homéomorphisme d'une surface. La question à laquelle nous intéressons est la suivante: Quels sont les automorphismes de $F_n$ qui sont géométriques?Nous donnons une réponse algorithmique pour la classe des automorphismes à croissance polynomiale (en s'autorisant à remplacer un automorphisme par une puissance).Pour cela, nous sommes amenés à étudier les automorphismes de graphes de groupes. En particulier, nous introduisons deux transformations élémentaires d'automorphismes de graphes de groupes: les quotients et les éclatements.Pour le cas particulier où l'automorphisme est un twist de Dehn partiel, on obtient un critère pour décider quand un tel twist de Dehn partiel est un véritable twist de Dehn.En appliquant le critère à plusieurs reprises sur un twist de Dehn cumulé, nous montrons que soit on peut ＂déplier＂ ce twist de Dehn cumulé jusqu'à obtenir un twist de Dehn ordinaire, soit que $phi$ est à croissance au moins quadratique (et par conséquent, n'est pas géométrique).Cela montre, au passage, que tout automorphisme du groupe libre à croissance linéaire admet une puissance qui est un twist de Dehn. Ce fait est connu des experts, et souvent utilisé, bien qu'il n'en existait pas de preuve formelle dans la littérature (à la connaissance de l'auteur).Pour conclure, on applique l'algorithme de Cohen-Lustig pour le transformer en twist de Dehn efficace, puis on applique l'algorithme Whitehead et des théorèmes classiques de Nielsen-Baer et Zieschang pour construire un modèle géométrique ou pour montrer qu'il n'est pas géométrique. / An automorphism $phi$ of a free group $F_n$ of finite rank $n geq 2$ is said to be geometric it is induced by a homeomorphism on a surface.In this thesis we concern ourselves with answering the question:Which precisely are the outer automorphisms of $F_n$ that are geometric?to which we give an algorithmical decision for the case of polynomially growing outer automorphisms, up to raising to certain positive power.In order to realize this algorithm, we establish the technique of quotient and blow-up automorphisms of graph-of-groups, which when apply for the special case of partial Dehn twist enables us to develop a criterion to decide whether the induced outer automorphism is an actual Dehn twist.Applying the criterion repeatedly on the special topological representative deriving from relative train track map, we are now able to either “unfold” this iterated relative Dehn twist representative level by level until eventually obtain an ordinary Dehn twist representative or show that $hat{phi}$ has at least quadratic growth hence is not geometric.As a side result, we also proved that every linearly growing automorphism of free group has a positive power which is a Dehn twist automorphism. This is a fact that has been taken for granted by many experts, although has no formal proof to be found in the literature.In the case of Dehn twist automorphisms, we then use the known algorithm to make the given Dehn twist representative efficient and apply the Whitehead algorithm as well as the classical theorems by Nielsen, Baers, Zieschangs and others to construct its geometric model or to show that it is not geometric.