X-Git-Url: http://gitweb.michael.orlitzky.com/?a=blobdiff_plain;f=mjo%2Fcone%2Fcone.py;h=132c6d9e8e63ebbc5e0065becd813f341e993dd0;hb=babaafed39f4919490631df5c5e9ba7339765cc0;hp=507b6cea5c24d0a4eb745d9245c85a1fbf4ddb90;hpb=494fe2e8517d40e3b71a54657e4a69a83cadf423;p=sage.d.git diff --git a/mjo/cone/cone.py b/mjo/cone/cone.py index 507b6ce..132c6d9 100644 --- a/mjo/cone/cone.py +++ b/mjo/cone/cone.py @@ -7,83 +7,198 @@ addsitedir(abspath('../../')) from sage.all import * - -def random_cone(min_dim=None, max_dim=None, min_rays=None, max_rays=None): +# TODO: This test fails, maybe due to a bug in the existing cone code. +# If we request enough generators to span the space, then the returned +# cone should equal the ambient space:: +# +# sage: K = random_cone(min_dim=5, max_dim=5, min_rays=10, max_rays=10) +# sage: K.lines().dimension() == K.lattice_dim() +# True + +def random_cone(min_dim=0, max_dim=None, min_rays=0, max_rays=None): r""" Generate a random rational convex polyhedral cone. Lower and upper bounds may be provided for both the dimension of the - ambient space and the number of generating rays of the cone. Any - parameters left unspecified will be chosen randomly. + ambient space and the number of generating rays of the cone. If a + lower bound is left unspecified, it defaults to zero. Unspecified + upper bounds will be chosen randomly. + + The number of generating rays is naturally limited to twice the + dimension of the ambient space. Take for example $\mathbb{R}^{2}$. + You could have the generators $\left\{ \pm e_{1}, \pm e_{2} + \right\}$, with cardinality $4 = 2 \cdot 2$; however any other ray + in the space is a nonnegative linear combination of those four. + + .. NOTE: + + If you do not explicitly request more than ``2 * max_dim`` rays, + a larger number may still be randomly generated. In that case, + the returned cone will simply be equal to the entire space. INPUT: - - ``min_dim`` (default: random) -- The minimum dimension of the ambient - lattice. + - ``min_dim`` (default: zero) -- A nonnegative integer representing the + minimum dimension of the ambient lattice. - - ``max_dim`` (default: random) -- The maximum dimension of the ambient + - ``max_dim`` (default: random) -- A nonnegative integer representing + the maximum dimension of the ambient lattice. - - ``min_rays`` (default: random) -- The minimum number of generating rays - of the cone. + - ``min_rays`` (default: zero) -- A nonnegative integer representing the + minimum number of generating rays of the + cone. - - ``max_rays`` (default: random) -- The maximum number of generating rays - of the cone. + - ``max_rays`` (default: random) -- A nonnegative integer representing the + maximum number of generating rays of + the cone. OUTPUT: A new, randomly generated cone. + A ``ValueError` will be thrown under the following conditions: + + * Any of ``min_dim``, ``max_dim``, ``min_rays``, or ``max_rays`` + are negative. + + * ``max_dim`` is less than ``min_dim``. + + * ``max_rays`` is less than ``min_rays``. + + * ``min_rays`` is greater than twice ``max_dim``. + + EXAMPLES: + + If we set the lower/upper bounds to zero, then our result is + predictable:: + + sage: random_cone(0,0,0,0) + 0-d cone in 0-d lattice N + + We can predict the dimension when ``min_dim == max_dim``:: + + sage: random_cone(min_dim=4, max_dim=4, min_rays=0, max_rays=0) + 0-d cone in 4-d lattice N + + Likewise for the number of rays when ``min_rays == max_rays``:: + + sage: random_cone(min_dim=10, max_dim=10, min_rays=10, max_rays=10) + 10-d cone in 10-d lattice N + TESTS: It's hard to test the output of a random process, but we can at least make sure that we get a cone back:: - sage: from sage.geometry.cone import is_Cone - sage: K = random_cone() - sage: is_Cone(K) # long time + sage: from sage.geometry.cone import is_Cone # long time + sage: K = random_cone() # long time + sage: is_Cone(K) # long time True + The upper/lower bounds are respected:: + + sage: K = random_cone(min_dim=5, max_dim=10, min_rays=3, max_rays=4) + sage: 5 <= K.lattice_dim() and K.lattice_dim() <= 10 + True + sage: 3 <= K.nrays() and K.nrays() <= 4 + True + + Ensure that an exception is raised when either lower bound is greater + than its respective upper bound:: + + sage: random_cone(min_dim=5, max_dim=2) + Traceback (most recent call last): + ... + ValueError: max_dim cannot be less than min_dim. + + sage: random_cone(min_rays=5, max_rays=2) + Traceback (most recent call last): + ... + ValueError: max_rays cannot be less than min_rays. + + And if we request too many rays:: + + sage: random_cone(min_rays=5, max_dim=1) + Traceback (most recent call last): + ... + ValueError: min_rays cannot be larger than twice max_dim. + """ + # Catch obvious mistakes so that we can generate clear error + # messages. + + if min_dim < 0: + raise ValueError('min_dim must be nonnegative.') + + if min_rays < 0: + raise ValueError('min_rays must be nonnegative.') + + if max_dim is not None: + if max_dim < 0: + raise ValueError('max_dim must be nonnegative.') + if (max_dim < min_dim): + raise ValueError('max_dim cannot be less than min_dim.') + if min_rays > 2*max_dim: + raise ValueError('min_rays cannot be larger than twice max_dim.') + + if max_rays is not None: + if max_rays < 0: + raise ValueError('max_rays must be nonnegative.') + if (max_rays < min_rays): + raise ValueError('max_rays cannot be less than min_rays.') + + def random_min_max(l,u): r""" - We need to handle four cases to prevent us from doing - something stupid like having an upper bound that's lower than - our lower bound. And we would need to repeat all of that logic - for the dimension/rays, so we consolidate it here. + We need to handle two cases for the upper bounds, and we need to do + the same thing for max_dim/max_rays. So we consolidate the logic here. """ - if l is None and u is None: - # They're both random, just return a random nonnegative - # integer. - return ZZ.random_element().abs() - - if l is not None and u is not None: - # Both were specified. Again, just make up a number and - # return it. If the user wants to give us u < l then he - # can have an exception. - return ZZ.random_element(l,u) - - if l is not None and u is None: - # In this case, we're generating the upper bound randomly - # GIVEN A LOWER BOUND. So we add a random nonnegative - # integer to the given lower bound. - u = l + ZZ.random_element().abs() - return ZZ.random_element(l,u) - - # Here we must be in the only remaining case, where we are - # given an upper bound but no lower bound. We might as well - # use zero. - return ZZ.random_element(0,u) + if u is None: + # The upper bound is unspecified; return a random integer + # in [l,infinity). + return l + ZZ.random_element().abs() + else: + # We have an upper bound, and it's greater than or equal + # to our lower bound. So we generate a random integer in + # [0,u-l], and then add it to l to get something in + # [l,u]. To understand the "+1", check the + # ZZ.random_element() docs. + return l + ZZ.random_element(u - l + 1) + d = random_min_max(min_dim, max_dim) r = random_min_max(min_rays, max_rays) L = ToricLattice(d) - rays = [L.random_element() for i in range(0,r)] - # We pass the lattice in case there are no rays. - return Cone(rays, lattice=L) + # The rays are trickier to generate, since we could generate v and + # 2*v as our "two rays." In that case, the resuting cone would + # have one generating ray. To avoid such a situation, we start by + # generating ``r`` rays where ``r`` is the number we want to end + # up with. + # + # However, since we're going to *check* whether or not we actually + # have ``r``, we need ``r`` rays to be attainable. So we need to + # limit ``r`` to twice the dimension of the ambient space. + # + r = min(r, 2*d) + rays = [L.random_element() for i in range(0, r)] + + # (The lattice parameter is required when no rays are given, so we + # pass it just in case ``r == 0``). + K = Cone(rays, lattice=L) + + # Now if we generated two of the "same" rays, we'll have fewer + # generating rays than ``r``. In that case, we keep making up new + # rays and recreating the cone until we get the right number of + # independent generators. + while r > K.nrays(): + rays.append(L.random_element()) + K = Cone(rays) + + return K def discrete_complementarity_set(K): @@ -140,7 +255,7 @@ def discrete_complementarity_set(K): The complementarity set of the dual can be obtained by switching the components of the complementarity set of the original cone:: - sage: K1 = random_cone(0,10,0,10) + sage: K1 = random_cone(max_dim=10, max_rays=10) sage: K2 = K1.dual() sage: expected = [(x,s) for (s,x) in discrete_complementarity_set(K2)] sage: actual = discrete_complementarity_set(K1) @@ -262,8 +377,8 @@ def lyapunov_rank(K): The Lyapunov rank should be additive on a product of cones:: - sage: K1 = random_cone(0,10,0,10) - sage: K2 = random_cone(0,10,0,10) + sage: K1 = random_cone(max_dim=10, max_rays=10) + sage: K2 = random_cone(max_dim=10, max_rays=10) sage: K = K1.cartesian_product(K2) sage: lyapunov_rank(K) == lyapunov_rank(K1) + lyapunov_rank(K2) True @@ -271,7 +386,7 @@ def lyapunov_rank(K): The dual cone `K^{*}` of ``K`` should have the same Lyapunov rank as ``K`` itself:: - sage: K = random_cone(0,10,0,10) + sage: K = random_cone(max_dim=10, max_rays=10) sage: lyapunov_rank(K) == lyapunov_rank(K.dual()) True