Index C++ Code Documentation math::minimize2 File: BASE/math/Minimize2.H Finds minimum of a function f:R^n-->R^1 minimize2_gradient() Finds minimum of a function f:R^n-->R^1 Adapted from Numerical Recipes in C, 2nd ed. sec. 10.6. (Gradient Method) function function to minimize dfunction gradient of 'function' computes the gradient of f at x y is function(x) report: after each line minimization, called with value report(y, x) minimize2_powell() Finds minimum of a function f:R^n-->R^1 Adapted from Numerical Recipes in C, 2nd ed. sec. 10.4. (Powell's Method) p: starting point xi initial directions an array of 'dimension' vectors of dimension 'dimension' iteration_maximum: maximum allowaAble iterations (e.g. 100) return value: 0: minimum found !=0: minimum not found public: minimize2( uint dimension ); constructor virtual ~minimize2(); destructor void set_precision_goal( double x ); double get_precision_goal(); void set_accuracy_goal( double x ); double get_accuracy_goal(); void set_iteration_limit( uint x ); uint get_iteration_limit(); uint get_iteration_count(); void set_gradient_dt( double x ); double get_gradient_dt(); void set_powell_dt( double x ); double get_powell_dt(); int find_minimum_gradient_auto( base::vector<double> &x, double &y ); minimize by the gradient method with an internally-provided gradient int find_minimum_gradient( base::vector<double> &x, double &y ); minimize by the gradient method with a user-provided gradient (not implemented) int find_minimum_powell( base::vector<double> &x, double &y ); minimize by powell's method int find_minimum_powell( base::vector<double> &x, base::vector<base::vector<double> > &dx, double &y ); minimize by powell's method (with dx) protected: virtual double function( base::vector<double> const &x ) = 0; function to minimize virtual void dfunction( base::vector<double> &, base::vector<double> const &, double); gradient of function to minimize virtual void report( double y, base::vector<double> const &x ); private: void dfunction_auto( base::vector<double> &gradient, base::vector<double> const &x, double y); int line_minimize( double &y, base::vector<double> &x, base::vector<double> &dx ); int bracket_minimum( double &ax, double &bx, double &cx, double &fa, double &fb, double &fc ); static inline void shift(double &a, double &b, double &c, double d); double _precision_goal; double _accuracy_goal; uint _iteration_limit; uint _iteration_count; double _gradient_dt; double _powell_dt; uint _dimension; base::vector<double> function_1dim_x; base::vector<double> function_1dim_dx; base::vector<double> function_1dim_tmp; class minimize1_subclass : public minimize1; minimize1_subclass minimize_1dim; friend class minimize1_subclass; static double const smallreal_double; minimize2( minimize2 const & ); Disabled copy constructor. minimize2 &operator=( minimize2 const & ); Disabled assignment operator.
	math::minimize2 GANG