random.h File Reference

Introduction

Definition of basic randomization functions.

See also

random.c.

Definition in file random.h.

Macros
#define	DEBUG_RANDOM   (_DEBUG>4?1:0)
	Debug the radom generation of numbers. More...

Functions
void	randomReset ()
	Resets the random seed. More...
void	randomSet (unsigned int seed)
	Sets the random seed. More...
double	randomDouble ()
	Returns a random double in the [0,1] interval. More...
double	randomNormal ()
	Returns a random double acording to a normal distribution. More...
double	randomInInterval (Tinterval *t)
	Returns a random double in the given interval. More...
unsigned int	randomMax (unsigned int m)
	Returns a random integer in the range [0,m]. More...
unsigned int	randomWithDistribution (unsigned int m, double s, double *d)
	Random number with a given discrete distribution. More...
void	randomOnBall (double r, unsigned int k, double *p)
	Random number on a k dimensional ball. More...
void	randomInBall (double r, unsigned int k, double *p)
	Random number in a k dimensional ball. More...
void	randomVector (unsigned int n, double *v)
	Random vector. More...

Macro Definition Documentation

DEBUG_RANDOM

#define DEBUG_RANDOM   (_DEBUG>4?1:0)

Used to debug the random generation of numbers.

Definition at line 12 of file random.h.

Function Documentation

randomReset()

void randomReset

(

)

Resets the random seed using the current time so, in principle, the reset is hard to predict.

Definition at line 18 of file random.c.

References randomSet().

Referenced by main().

randomSet()

void randomSet

(

unsigned int

seed

)

Sets the random to the given value. This is used to get repetitive sequences of random numbers in order to debug mainly.

Please, note that when using advanced numerical libraries (BLAS) the result of numerical computations can be different at each run (https://bebop.cs.berkeley.edu/reproblas). The differences are tiny (at the machine precision) but in some systems (specially with dynamics) such errors can accumulate and produce a different result in each run.

To avoid this, you can increase the DEBUG level (in the main CMake file) since then BLAS is not used. However, parallelism will not be used either and the execution will be much slower.

Parameters

seed	The new random seed.

Definition at line 23 of file random.c.

Referenced by main(), and randomReset().

randomDouble()

double randomDouble

(

)

Returns a random number in the [0,1] interval selected with a uniform distribution.

Returns

A random double in [0,1].

Definition at line 35 of file random.c.

Referenced by AddBranchToAtlasDynamicRRT(), ApproachState(), AtlasRRT(), AtlasRRTSample(), AtlasRRTValidateSample(), BuildAtlasFromPoint(), GenerateProjectionPlanes(), GetActionFromID(), kinoEST(), randomInBall(), randomInInterval(), randomMax(), randomNormal(), randomVector(), randomWithDistribution(), RRTSample(), SimplexAddNewConstraint(), and TransitionTestRRT().

randomNormal()

double randomNormal

(

)

Returns a random number following a normal distribution with mean 0 and standard deviation 1.

Returns

A random double.

Definition at line 46 of file random.c.

References FALSE, randomDouble(), and TRUE.

Referenced by randomInBall(), and randomOnBall().

randomInInterval()

double randomInInterval

(

Tinterval *

t

)

Returns a random number in the given interval selected with a uniform distribution.

Parameters

t	The interval where to drawn the random sample.

Returns

A random double in the given interval.

Definition at line 83 of file random.c.

References LowerLimit(), randomDouble(), and UpperLimit().

Referenced by CuikGradientInBox(), CuikNewtonInBox(), IncrementalSampleCuikSystemInBox(), RandomPointInBox(), RandomPointInPolytope(), and SampleCuikSystemInBox().

randomMax()

unsigned int randomMax

(

unsigned int

m

)

Returns a random integer in the range [0,m] selected with a uniform distribution. Observe that m is included as a possible outcome.

Parameters

m	The upper limit for the random generator.

Returns

The randomly selected integer in [0,m].

Definition at line 99 of file random.c.

References randomDouble().

Referenced by AddBranchToAtlasDynamicRRT(), ComputeSplitDimInt(), IncrementalSampleCuikSystemInBox(), kinoEST(), PolytopeBoundaryPointFromExternalCorner(), RandomPointInAtlas(), RandomPointInAtlasTree(), RandomSmooth(), and SmoothPathInAtlasRRT().

randomWithDistribution()

unsigned int randomWithDistribution	(	unsigned int	m,
		double	s,
		double *	d
	)

Gets a random number following an arbitrary random discrete distribution.

Parameters

m	The size of the distribution (number of options).
s	Sum of the values in d. If zero the sum is computed inside this function.
d	The value for each option. The values in d are assume possitive but not normalized.

Definition at line 118 of file random.c.

References randomDouble().

Referenced by RandomPointInAtlas(), and RandomPointInAtlasTree().

randomOnBall()

void randomOnBall	(	double	r,
		unsigned int	k,
		double *	p
	)

Selects a random point on the surface of a k dimensional ball with uniform distribution.

Parameters

r	Radius of the ball.
k	Dimension of the space where the ball is defiend.
p	Space for the output point. Must be allocated externally.

Definition at line 146 of file random.c.

References randomNormal().

Referenced by PolytopeRandomPointOnBoundary(), and SPolytopeRandomPointOnBoundary().

randomInBall()

void randomInBall	(	double	r,
		unsigned int	k,
		double *	p
	)

Selects a random point inside a k dimensional ball with uniform distribution.

Parameters

r	Radius of the ball.
k	Dimension of the space where the ball is defiend.
p	Space for the output point. Must be allocated externally.

Definition at line 166 of file random.c.

References randomDouble(), and randomNormal().

Referenced by RandomPointInSPolytope(), and SPolytopeVolume().

randomVector()

void randomVector	(	unsigned int	n,
		double *	v
	)

Random vector with numbers uniformly distributed in the [0,1] interval.

Parameters

n	The size fo the vector.
v	The vector to fill.

Definition at line 189 of file random.c.

References randomDouble().