A. We carry out an in-depth study of Martin compactifications of affine buildings, that is, from the viewpoint of potential theory and random walks. This work does not use any group action on buildings, although in the end all the results can be stated within the framework of the Bruhat-Tits theory of semisimple groups over non-Archimedean local fields. This choice should allow the use of these building compactifications in intriguing geometric group theory situations, where only lattice actions are available. The choice avoiding the use of a group action requires going back to certain more classical compactification procedures. The compactified spaces we obtain use and, at the same time, make it possible to understand geometrically the descriptions of asymptotic behavior of kernels resulting from the non-Archimedean harmonic analysis on affine buildings.
IThis paper deals with the Martin compactifications of affine buildings. In other words, it makes a connection between two very different mathematical topics. On the one hand, affine buildings are relevant to algebra and geometry and, on the other hand, Martin compactifications refer to analysis, more precisely potential and probability theory. Therefore, our first task in this introduction, before mentioning the new results, is to introduce these two fields independently but in a way making them compatible with one another. At this stage let us simply say that dealing with compactifications associated to potential theory allows us to construct, from a concrete viewpoint, compactifications that before this approach could only be obtained artificially. Conversely, these compactifications provide a geometric way of understanding the various factors in the asymptotic formula for the Green function obtained previously by Gelfand-Fourier analytic methods.In what follows, the geometry on which the various analytic concepts are defined (such as random walks, or heat and Martin kernels) are affine buildings. In many situations, the latter spaces are non-Archimedean analogues of Riemannian symmetric spaces; they were indeed designed for this purpose by F. Bruhat and J. Tits (see [13] and [14] for the full theory, and [51] for an overview). Affine buildings thus provide the well-adapted geometry that enables one to understand semisimple algebraic groups over non-Archimedean local fields, such as classical matrix groups over finite extensions of the field of -adic numbers Q (e.g. the group SL (Q ) itself). However, some affine buildings of low rank do not come from algebraic groups, and thus have interesting features in geometric group theory. For this reason we avoid using group actions in the paper, even though we are led by this analogy. This approach is completely parallel to the way a semisimple real Lie group without compact factor is understood, that is via its action on the associated symmetric space = / where is a maximal compact subgroup (see [25] and [37]). As a result, geometric proofs of crucial tools in Lie theory and in representation theory, such as the we...