Math::LinearCombination - sum of variables with a numerical coefficient
use Math::LinearCombination; use Math::SimpleVariable; # for the variable objects # build a linear combination my $x1 = new Math::SimpleVariable(name => 'x1'); my $x2 = new Math::SimpleVariable(name => 'x2'); my $lc = new Math::LinearCombination(); $lc->add_entry(var => $x1, coeff => 3.0); $lc->add_entry(var => $x2, coeff => 1.7); $lc->add_entry(var => $x2, coeff => 0.3); # so x2 has a coefficient of 2.0 print $lc->stringify(), "\n"; # do some manipulations $lc->negate_inplace(); # reverts the coefficient signs $lc->multiply_with_constant_inplace(2.0); # doubles all coefficients # evaluate the linear combination $x1->{value} = 3; $x2->{value} = -1; print $lc->evaluate(), "\n"; # prints -14
Math::LinearCombination is a module for representing mathematical linear combinations of variables, i.e. expressions of the format
a1 * x1 + a2 * x2 + ... + an * xn
with x1, x2, ..., xn variables, and a1, a2, ..., an numerical coefficients. Evaluation and manipulation of linear combinations is also supported. The numerical coefficients a_i and variables x_i are stored as pairs in an internal data structure and should not be manipulated directly. All access and manipulation should be performed through the methods.
It is important to note that no specific class is required for the variable objects. You can provide objects of any class, provided that the following methods are defined on those objects:
returning a string with the variable name.
returning a unique identifier for that variable. For most applications it will suffice to have id() invoke name().
returning a numerical evaluation of the variable.
The Math::LinearCombination class was designed together with Math::SimpleVariable. The latter supports all the required methods, and it is thus logical to use only Math::SimpleVariable objects in your linear combinations, or brew your own class which is derived from Math::SimpleVariable.
The following methods are available for Math::LinearCombination objects:
constructs a new Math::SimpleVariable object. An existing Math::SimpleVariable object can be passed to it optionally, in which case a clone of that object is returned. (see also the clone() method).
also constructs a new Math::SimpleVariable object, but additionally initializes it with variable-coefficient pairs x_i, a_i. The number of arguments should thus be even, and the variable objects need to obey the requirements imposed on variables.
returns an exact copy of $lc, with none of the variables or coefficients shared. I.e. you can change the contents of the new (old) linear combination without any impact on the old (new) one.
adds the variable $x to the linear combination with $a as coefficient. add_entry() throws an error when a variable with the same id() is already present in the linear combination. So do not use add_entry() for adding linear combinations, but use add_inplace() or the '+' operator (see below) instead.
returns a ref to a hash with all the entries of the linear combination. The hash is sorted on the id() of the variables, and each entry is a ref to a hash with the following fields:
returns an array with all the variable objects in the linear combination. get_variables() is context aware, so you can invoke it as
$ra_vars = $lc->get_variables
to return a reference to the array of variables instead. The variables are sorted on their id().
same as get_variables(), but returns the coefficients instead. The coefficients are also sorted on the id() of the corresponding variables.
mathematically adds the Math::LinearCombination object $lc2 to $lc and returns the changed $lc.
mathematically adds $lc2 to $lc. The result is stored in a newly created Math::LinearCombination object. Both $lc and $lc2 are left in the same state as before.
inverts the sign of the coefficient for each variable in the linear combination.
mathematically subtracts $lc2 from $lc. The results is stored in a newly created Math::LinearCombination object. Both $lc and $lc2 are left in the same state as before.
multiplies each coefficient in $lc with the numerical constant $c.
mathematically multiplies $lc with $lc2. The result is stored in a newly created Math::LinearCombination object. Both $lc and $lc2 are left in the same state as before.
divides $lc by the numerical constant $c. The result is stored in a newly created Math::LinearCombination object. $lc is left in the same state as before. Note that it is not possible to divide a linear combination by another linear combination, as the result is generally not a linear combination.
evaluates the linear combination $lc numerically, using the values of the variables obtained by invoking evaluate() on them.
removes all variable-coefficient pairs with zero coefficients from the linear combination. This method is used internally in the methods above, so normally you should never need it.
returns a string representing the linear combination. Returns '0.0' for empty linear combinations.
In order to make the mathematical manipulation of linear combinations less verbose, a number of operators have been overloaded to use the methods above. The overloaded operators are:
This is CVS $Revision: 1.11 $ of Math::LinearCombination, last edited at $Date: 2001/10/31 12:50:02 $.
Wim Verhaegen <wimv@cpan.org>
Copyright (C) 2001 Wim Verhaegen. All rights reserved. This program is free software; you may redistribute and/or modify it under the same terms as Perl itself.