Ws. Liu et Hj. Sussmann, SHORTEST PATHS FOR SUB-RIEMANNIAN METRICS ON RANK-2 DISTRIBUTIONS - INTRODUCTION, Memoirs of the American Mathematical Society, 118(564), 1995, pp. 1
We study length-minimizing arcs in sub-Riemannian manifolds (M, E, G)
whose metric G is defined on a rank-two bracket-generating distributio
n E. It is well known that all length-minimizing arcs are extrema ls,
and that these extremals are either ''normal'' or ''abnormal.'' Normal
extremals are locally optimal, in the sense that every sufficiently s
hort piece of such an extremal is a minimizer. The question whether ev
ery length-minimizer is a normal extremal remained open for several ye
ars, and was recently settled by R. Montgomery, who exhibited a counte
rexample. But Montgomery's geometric optimality proof depends heavily
on special properties of his example and still leaves open the questio
n whether abnormal minimizers are an exceptional phenomenon or a commo
n occurrence. We present an analytic technique for proving local optim
ality of a large class of abnormal extremals that we call ''regular.''
Our technique is based on (a) a ''normal form theorem,'' stating that
, locally, a regular abnormal extremal can always be put in a special
form by a suitable change of coordinates, and (b) an inequality showin
g that, once a trajectory is in this special form? then local optimali
ty follows. Using this approach we prove that regular abnormal extrema
ls are locally optimal. If E satisfies a mild additional restriction -
valid in particular for all regular 2-dimensional distributions and f
or generic 2-dimensional distributions - then regular abnormal extrema
ls are ''typical'' (in a sense made precise in the text), so our resul
t implies that the abnormal minimizers are ubiquitous rather than exce
ptional. We also discuss some related issues, and in particular show,
by means of an example, that a smooth abnormal extremal need not be lo
cally optimal, even if in addition it belongs to the class - recently
studied by Bryant and Hsu - of C-1-rigid curves.