470 lines
		
	
	
		
			16 KiB
		
	
	
	
		
			Plaintext
		
	
	
	
	
	
			
		
		
	
	
			470 lines
		
	
	
		
			16 KiB
		
	
	
	
		
			Plaintext
		
	
	
	
	
	
| // Copyright John Maddock 2012.
 | |
| // Use, modification and distribution are subject to the
 | |
| // Boost Software License, Version 1.0.
 | |
| // (See accompanying file LICENSE_1_0.txt
 | |
| // or copy at http://www.boost.org/LICENSE_1_0.txt)
 | |
| 
 | |
| #ifndef BOOST_MATH_AIRY_HPP
 | |
| #define BOOST_MATH_AIRY_HPP
 | |
| 
 | |
| #include <limits>
 | |
| #include <boost/math/special_functions/math_fwd.hpp>
 | |
| #include <boost/math/special_functions/bessel.hpp>
 | |
| #include <boost/math/special_functions/cbrt.hpp>
 | |
| #include <boost/math/special_functions/detail/airy_ai_bi_zero.hpp>
 | |
| #include <boost/math/tools/roots.hpp>
 | |
| 
 | |
| namespace boost{ namespace math{
 | |
| 
 | |
| namespace detail{
 | |
| 
 | |
| template <class T, class Policy>
 | |
| T airy_ai_imp(T x, const Policy& pol)
 | |
| {
 | |
|    BOOST_MATH_STD_USING
 | |
| 
 | |
|    if(x < 0)
 | |
|    {
 | |
|       T p = (-x * sqrt(-x) * 2) / 3;
 | |
|       T v = T(1) / 3;
 | |
|       T j1 = boost::math::cyl_bessel_j(v, p, pol);
 | |
|       T j2 = boost::math::cyl_bessel_j(-v, p, pol);
 | |
|       T ai = sqrt(-x) * (j1 + j2) / 3;
 | |
|       //T bi = sqrt(-x / 3) * (j2 - j1);
 | |
|       return ai;
 | |
|    }
 | |
|    else if(fabs(x * x * x) / 6 < tools::epsilon<T>())
 | |
|    {
 | |
|       T tg = boost::math::tgamma(constants::twothirds<T>(), pol);
 | |
|       T ai = 1 / (pow(T(3), constants::twothirds<T>()) * tg);
 | |
|       //T bi = 1 / (sqrt(boost::math::cbrt(T(3))) * tg);
 | |
|       return ai;
 | |
|    }
 | |
|    else
 | |
|    {
 | |
|       T p = 2 * x * sqrt(x) / 3;
 | |
|       T v = T(1) / 3;
 | |
|       //T j1 = boost::math::cyl_bessel_i(-v, p, pol);
 | |
|       //T j2 = boost::math::cyl_bessel_i(v, p, pol);
 | |
|       //
 | |
|       // Note that although we can calculate ai from j1 and j2, the accuracy is horrible
 | |
|       // as we're subtracting two very large values, so use the Bessel K relation instead:
 | |
|       //
 | |
|       T ai = cyl_bessel_k(v, p, pol) * sqrt(x / 3) / boost::math::constants::pi<T>();  //sqrt(x) * (j1 - j2) / 3;
 | |
|       //T bi = sqrt(x / 3) * (j1 + j2);
 | |
|       return ai;
 | |
|    }
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| T airy_bi_imp(T x, const Policy& pol)
 | |
| {
 | |
|    BOOST_MATH_STD_USING
 | |
| 
 | |
|    if(x < 0)
 | |
|    {
 | |
|       T p = (-x * sqrt(-x) * 2) / 3;
 | |
|       T v = T(1) / 3;
 | |
|       T j1 = boost::math::cyl_bessel_j(v, p, pol);
 | |
|       T j2 = boost::math::cyl_bessel_j(-v, p, pol);
 | |
|       //T ai = sqrt(-x) * (j1 + j2) / 3;
 | |
|       T bi = sqrt(-x / 3) * (j2 - j1);
 | |
|       return bi;
 | |
|    }
 | |
|    else if(fabs(x * x * x) / 6 < tools::epsilon<T>())
 | |
|    {
 | |
|       T tg = boost::math::tgamma(constants::twothirds<T>(), pol);
 | |
|       //T ai = 1 / (pow(T(3), constants::twothirds<T>()) * tg);
 | |
|       T bi = 1 / (sqrt(boost::math::cbrt(T(3))) * tg);
 | |
|       return bi;
 | |
|    }
 | |
|    else
 | |
|    {
 | |
|       T p = 2 * x * sqrt(x) / 3;
 | |
|       T v = T(1) / 3;
 | |
|       T j1 = boost::math::cyl_bessel_i(-v, p, pol);
 | |
|       T j2 = boost::math::cyl_bessel_i(v, p, pol);
 | |
|       T bi = sqrt(x / 3) * (j1 + j2);
 | |
|       return bi;
 | |
|    }
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| T airy_ai_prime_imp(T x, const Policy& pol)
 | |
| {
 | |
|    BOOST_MATH_STD_USING
 | |
| 
 | |
|    if(x < 0)
 | |
|    {
 | |
|       T p = (-x * sqrt(-x) * 2) / 3;
 | |
|       T v = T(2) / 3;
 | |
|       T j1 = boost::math::cyl_bessel_j(v, p, pol);
 | |
|       T j2 = boost::math::cyl_bessel_j(-v, p, pol);
 | |
|       T aip = -x * (j1 - j2) / 3;
 | |
|       return aip;
 | |
|    }
 | |
|    else if(fabs(x * x) / 2 < tools::epsilon<T>())
 | |
|    {
 | |
|       T tg = boost::math::tgamma(constants::third<T>(), pol);
 | |
|       T aip = 1 / (boost::math::cbrt(T(3)) * tg);
 | |
|       return -aip;
 | |
|    }
 | |
|    else
 | |
|    {
 | |
|       T p = 2 * x * sqrt(x) / 3;
 | |
|       T v = T(2) / 3;
 | |
|       //T j1 = boost::math::cyl_bessel_i(-v, p, pol);
 | |
|       //T j2 = boost::math::cyl_bessel_i(v, p, pol);
 | |
|       //
 | |
|       // Note that although we can calculate ai from j1 and j2, the accuracy is horrible
 | |
|       // as we're subtracting two very large values, so use the Bessel K relation instead:
 | |
|       //
 | |
|       T aip = -cyl_bessel_k(v, p, pol) * x / (boost::math::constants::root_three<T>() * boost::math::constants::pi<T>());
 | |
|       return aip;
 | |
|    }
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| T airy_bi_prime_imp(T x, const Policy& pol)
 | |
| {
 | |
|    BOOST_MATH_STD_USING
 | |
| 
 | |
|    if(x < 0)
 | |
|    {
 | |
|       T p = (-x * sqrt(-x) * 2) / 3;
 | |
|       T v = T(2) / 3;
 | |
|       T j1 = boost::math::cyl_bessel_j(v, p, pol);
 | |
|       T j2 = boost::math::cyl_bessel_j(-v, p, pol);
 | |
|       T aip = -x * (j1 + j2) / constants::root_three<T>();
 | |
|       return aip;
 | |
|    }
 | |
|    else if(fabs(x * x) / 2 < tools::epsilon<T>())
 | |
|    {
 | |
|       T tg = boost::math::tgamma(constants::third<T>(), pol);
 | |
|       T bip = sqrt(boost::math::cbrt(T(3))) / tg;
 | |
|       return bip;
 | |
|    }
 | |
|    else
 | |
|    {
 | |
|       T p = 2 * x * sqrt(x) / 3;
 | |
|       T v = T(2) / 3;
 | |
|       T j1 = boost::math::cyl_bessel_i(-v, p, pol);
 | |
|       T j2 = boost::math::cyl_bessel_i(v, p, pol);
 | |
|       T aip = x * (j1 + j2) / boost::math::constants::root_three<T>();
 | |
|       return aip;
 | |
|    }
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| T airy_ai_zero_imp(int m, const Policy& pol)
 | |
| {
 | |
|    BOOST_MATH_STD_USING // ADL of std names, needed for log, sqrt.
 | |
| 
 | |
|    // Handle cases when a negative zero (negative rank) is requested.
 | |
|    if(m < 0)
 | |
|    {
 | |
|       return policies::raise_domain_error<T>("boost::math::airy_ai_zero<%1%>(%1%, int)",
 | |
|          "Requested the %1%'th zero, but the rank must be 1 or more !", static_cast<T>(m), pol);
 | |
|    }
 | |
| 
 | |
|    // Handle case when the zero'th zero is requested.
 | |
|    if(m == 0U)
 | |
|    {
 | |
|       return policies::raise_domain_error<T>("boost::math::airy_ai_zero<%1%>(%1%,%1%)",
 | |
|         "The requested rank of the zero is %1%, but must be 1 or more !", static_cast<T>(m), pol);
 | |
|    }
 | |
| 
 | |
|    // Set up the initial guess for the upcoming root-finding.
 | |
|    const T guess_root = boost::math::detail::airy_zero::airy_ai_zero_detail::initial_guess<T>(m);
 | |
| 
 | |
|    // Select the maximum allowed iterations based on the number
 | |
|    // of decimal digits in the numeric type T, being at least 12.
 | |
|    const int my_digits10 = static_cast<int>(static_cast<float>(policies::digits<T, Policy>() * 0.301F));
 | |
| 
 | |
|    const boost::uintmax_t iterations_allowed = static_cast<boost::uintmax_t>((std::max)(12, my_digits10 * 2));
 | |
| 
 | |
|    boost::uintmax_t iterations_used = iterations_allowed;
 | |
| 
 | |
|    // Use a dynamic tolerance because the roots get closer the higher m gets.
 | |
|    T tolerance;
 | |
| 
 | |
|    if     (m <=   10) { tolerance = T(0.3F); }
 | |
|    else if(m <=  100) { tolerance = T(0.1F); }
 | |
|    else if(m <= 1000) { tolerance = T(0.05F); }
 | |
|    else               { tolerance = T(1) / sqrt(T(m)); }
 | |
| 
 | |
|    // Perform the root-finding using Newton-Raphson iteration from Boost.Math.
 | |
|    const T am =
 | |
|       boost::math::tools::newton_raphson_iterate(
 | |
|          boost::math::detail::airy_zero::airy_ai_zero_detail::function_object_ai_and_ai_prime<T, Policy>(pol),
 | |
|          guess_root,
 | |
|          T(guess_root - tolerance),
 | |
|          T(guess_root + tolerance),
 | |
|          policies::digits<T, Policy>(),
 | |
|          iterations_used);
 | |
| 
 | |
|    static_cast<void>(iterations_used);
 | |
| 
 | |
|    return am;
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| T airy_bi_zero_imp(int m, const Policy& pol)
 | |
| {
 | |
|    BOOST_MATH_STD_USING // ADL of std names, needed for log, sqrt.
 | |
| 
 | |
|    // Handle cases when a negative zero (negative rank) is requested.
 | |
|    if(m < 0)
 | |
|    {
 | |
|       return policies::raise_domain_error<T>("boost::math::airy_bi_zero<%1%>(%1%, int)",
 | |
|          "Requested the %1%'th zero, but the rank must 1 or more !", static_cast<T>(m), pol);
 | |
|    }
 | |
| 
 | |
|    // Handle case when the zero'th zero is requested.
 | |
|    if(m == 0U)
 | |
|    {
 | |
|       return policies::raise_domain_error<T>("boost::math::airy_bi_zero<%1%>(%1%,%1%)",
 | |
|         "The requested rank of the zero is %1%, but must be 1 or more !", static_cast<T>(m), pol);
 | |
|    }
 | |
|    // Set up the initial guess for the upcoming root-finding.
 | |
|    const T guess_root = boost::math::detail::airy_zero::airy_bi_zero_detail::initial_guess<T>(m);
 | |
| 
 | |
|    // Select the maximum allowed iterations based on the number
 | |
|    // of decimal digits in the numeric type T, being at least 12.
 | |
|    const int my_digits10 = static_cast<int>(static_cast<float>(policies::digits<T, Policy>() * 0.301F));
 | |
| 
 | |
|    const boost::uintmax_t iterations_allowed = static_cast<boost::uintmax_t>((std::max)(12, my_digits10 * 2));
 | |
| 
 | |
|    boost::uintmax_t iterations_used = iterations_allowed;
 | |
| 
 | |
|    // Use a dynamic tolerance because the roots get closer the higher m gets.
 | |
|    T tolerance;
 | |
| 
 | |
|    if     (m <=   10) { tolerance = T(0.3F); }
 | |
|    else if(m <=  100) { tolerance = T(0.1F); }
 | |
|    else if(m <= 1000) { tolerance = T(0.05F); }
 | |
|    else               { tolerance = T(1) / sqrt(T(m)); }
 | |
| 
 | |
|    // Perform the root-finding using Newton-Raphson iteration from Boost.Math.
 | |
|    const T bm =
 | |
|       boost::math::tools::newton_raphson_iterate(
 | |
|          boost::math::detail::airy_zero::airy_bi_zero_detail::function_object_bi_and_bi_prime<T, Policy>(pol),
 | |
|          guess_root,
 | |
|          T(guess_root - tolerance),
 | |
|          T(guess_root + tolerance),
 | |
|          policies::digits<T, Policy>(),
 | |
|          iterations_used);
 | |
| 
 | |
|    static_cast<void>(iterations_used);
 | |
| 
 | |
|    return bm;
 | |
| }
 | |
| 
 | |
| } // namespace detail
 | |
| 
 | |
| template <class T, class Policy>
 | |
| inline typename tools::promote_args<T>::type airy_ai(T x, const Policy&)
 | |
| {
 | |
|    BOOST_FPU_EXCEPTION_GUARD
 | |
|    typedef typename tools::promote_args<T>::type result_type;
 | |
|    typedef typename policies::evaluation<result_type, Policy>::type value_type;
 | |
|    typedef typename policies::normalise<
 | |
|       Policy, 
 | |
|       policies::promote_float<false>, 
 | |
|       policies::promote_double<false>, 
 | |
|       policies::discrete_quantile<>,
 | |
|       policies::assert_undefined<> >::type forwarding_policy;
 | |
| 
 | |
|    return policies::checked_narrowing_cast<result_type, Policy>(detail::airy_ai_imp<value_type>(static_cast<value_type>(x), forwarding_policy()), "boost::math::airy<%1%>(%1%)");
 | |
| }
 | |
| 
 | |
| template <class T>
 | |
| inline typename tools::promote_args<T>::type airy_ai(T x)
 | |
| {
 | |
|    return airy_ai(x, policies::policy<>());
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| inline typename tools::promote_args<T>::type airy_bi(T x, const Policy&)
 | |
| {
 | |
|    BOOST_FPU_EXCEPTION_GUARD
 | |
|    typedef typename tools::promote_args<T>::type result_type;
 | |
|    typedef typename policies::evaluation<result_type, Policy>::type value_type;
 | |
|    typedef typename policies::normalise<
 | |
|       Policy, 
 | |
|       policies::promote_float<false>, 
 | |
|       policies::promote_double<false>, 
 | |
|       policies::discrete_quantile<>,
 | |
|       policies::assert_undefined<> >::type forwarding_policy;
 | |
| 
 | |
|    return policies::checked_narrowing_cast<result_type, Policy>(detail::airy_bi_imp<value_type>(static_cast<value_type>(x), forwarding_policy()), "boost::math::airy<%1%>(%1%)");
 | |
| }
 | |
| 
 | |
| template <class T>
 | |
| inline typename tools::promote_args<T>::type airy_bi(T x)
 | |
| {
 | |
|    return airy_bi(x, policies::policy<>());
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| inline typename tools::promote_args<T>::type airy_ai_prime(T x, const Policy&)
 | |
| {
 | |
|    BOOST_FPU_EXCEPTION_GUARD
 | |
|    typedef typename tools::promote_args<T>::type result_type;
 | |
|    typedef typename policies::evaluation<result_type, Policy>::type value_type;
 | |
|    typedef typename policies::normalise<
 | |
|       Policy, 
 | |
|       policies::promote_float<false>, 
 | |
|       policies::promote_double<false>, 
 | |
|       policies::discrete_quantile<>,
 | |
|       policies::assert_undefined<> >::type forwarding_policy;
 | |
| 
 | |
|    return policies::checked_narrowing_cast<result_type, Policy>(detail::airy_ai_prime_imp<value_type>(static_cast<value_type>(x), forwarding_policy()), "boost::math::airy<%1%>(%1%)");
 | |
| }
 | |
| 
 | |
| template <class T>
 | |
| inline typename tools::promote_args<T>::type airy_ai_prime(T x)
 | |
| {
 | |
|    return airy_ai_prime(x, policies::policy<>());
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| inline typename tools::promote_args<T>::type airy_bi_prime(T x, const Policy&)
 | |
| {
 | |
|    BOOST_FPU_EXCEPTION_GUARD
 | |
|    typedef typename tools::promote_args<T>::type result_type;
 | |
|    typedef typename policies::evaluation<result_type, Policy>::type value_type;
 | |
|    typedef typename policies::normalise<
 | |
|       Policy, 
 | |
|       policies::promote_float<false>, 
 | |
|       policies::promote_double<false>, 
 | |
|       policies::discrete_quantile<>,
 | |
|       policies::assert_undefined<> >::type forwarding_policy;
 | |
| 
 | |
|    return policies::checked_narrowing_cast<result_type, Policy>(detail::airy_bi_prime_imp<value_type>(static_cast<value_type>(x), forwarding_policy()), "boost::math::airy<%1%>(%1%)");
 | |
| }
 | |
| 
 | |
| template <class T>
 | |
| inline typename tools::promote_args<T>::type airy_bi_prime(T x)
 | |
| {
 | |
|    return airy_bi_prime(x, policies::policy<>());
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| inline T airy_ai_zero(int m, const Policy& /*pol*/)
 | |
| {
 | |
|    BOOST_FPU_EXCEPTION_GUARD
 | |
|    typedef typename policies::evaluation<T, Policy>::type value_type;
 | |
|    typedef typename policies::normalise<
 | |
|       Policy, 
 | |
|       policies::promote_float<false>, 
 | |
|       policies::promote_double<false>, 
 | |
|       policies::discrete_quantile<>,
 | |
|       policies::assert_undefined<> >::type forwarding_policy;
 | |
| 
 | |
|    BOOST_STATIC_ASSERT_MSG(    false == std::numeric_limits<T>::is_specialized
 | |
|                            || (   true  == std::numeric_limits<T>::is_specialized
 | |
|                                && false == std::numeric_limits<T>::is_integer),
 | |
|                            "Airy value type must be a floating-point type.");
 | |
| 
 | |
|    return policies::checked_narrowing_cast<T, Policy>(detail::airy_ai_zero_imp<value_type>(m, forwarding_policy()), "boost::math::airy_ai_zero<%1%>(unsigned)");
 | |
| }
 | |
| 
 | |
| template <class T>
 | |
| inline T airy_ai_zero(int m)
 | |
| {
 | |
|    return airy_ai_zero<T>(m, policies::policy<>());
 | |
| }
 | |
| 
 | |
| template <class T, class OutputIterator, class Policy>
 | |
| inline OutputIterator airy_ai_zero(
 | |
|                          int start_index,
 | |
|                          unsigned number_of_zeros,
 | |
|                          OutputIterator out_it,
 | |
|                          const Policy& pol)
 | |
| {
 | |
|    typedef T result_type;
 | |
| 
 | |
|    BOOST_STATIC_ASSERT_MSG(    false == std::numeric_limits<T>::is_specialized
 | |
|                            || (   true  == std::numeric_limits<T>::is_specialized
 | |
|                                && false == std::numeric_limits<T>::is_integer),
 | |
|                            "Airy value type must be a floating-point type.");
 | |
| 
 | |
|    for(unsigned i = 0; i < number_of_zeros; ++i)
 | |
|    {
 | |
|       *out_it = boost::math::airy_ai_zero<result_type>(start_index + i, pol);
 | |
|       ++out_it;
 | |
|    }
 | |
|    return out_it;
 | |
| }
 | |
| 
 | |
| template <class T, class OutputIterator>
 | |
| inline OutputIterator airy_ai_zero(
 | |
|                          int start_index,
 | |
|                          unsigned number_of_zeros,
 | |
|                          OutputIterator out_it)
 | |
| {
 | |
|    return airy_ai_zero<T>(start_index, number_of_zeros, out_it, policies::policy<>());
 | |
| }
 | |
| 
 | |
| template <class T, class Policy>
 | |
| inline T airy_bi_zero(int m, const Policy& /*pol*/)
 | |
| {
 | |
|    BOOST_FPU_EXCEPTION_GUARD
 | |
|    typedef typename policies::evaluation<T, Policy>::type value_type;
 | |
|    typedef typename policies::normalise<
 | |
|       Policy, 
 | |
|       policies::promote_float<false>, 
 | |
|       policies::promote_double<false>, 
 | |
|       policies::discrete_quantile<>,
 | |
|       policies::assert_undefined<> >::type forwarding_policy;
 | |
| 
 | |
|    BOOST_STATIC_ASSERT_MSG(    false == std::numeric_limits<T>::is_specialized
 | |
|                            || (   true  == std::numeric_limits<T>::is_specialized
 | |
|                                && false == std::numeric_limits<T>::is_integer),
 | |
|                            "Airy value type must be a floating-point type.");
 | |
| 
 | |
|    return policies::checked_narrowing_cast<T, Policy>(detail::airy_bi_zero_imp<value_type>(m, forwarding_policy()), "boost::math::airy_bi_zero<%1%>(unsigned)");
 | |
| }
 | |
| 
 | |
| template <typename T>
 | |
| inline T airy_bi_zero(int m)
 | |
| {
 | |
|    return airy_bi_zero<T>(m, policies::policy<>());
 | |
| }
 | |
| 
 | |
| template <class T, class OutputIterator, class Policy>
 | |
| inline OutputIterator airy_bi_zero(
 | |
|                          int start_index,
 | |
|                          unsigned number_of_zeros,
 | |
|                          OutputIterator out_it,
 | |
|                          const Policy& pol)
 | |
| {
 | |
|    typedef T result_type;
 | |
| 
 | |
|    BOOST_STATIC_ASSERT_MSG(    false == std::numeric_limits<T>::is_specialized
 | |
|                            || (   true  == std::numeric_limits<T>::is_specialized
 | |
|                                && false == std::numeric_limits<T>::is_integer),
 | |
|                            "Airy value type must be a floating-point type.");
 | |
| 
 | |
|    for(unsigned i = 0; i < number_of_zeros; ++i)
 | |
|    {
 | |
|       *out_it = boost::math::airy_bi_zero<result_type>(start_index + i, pol);
 | |
|       ++out_it;
 | |
|    }
 | |
|    return out_it;
 | |
| }
 | |
| 
 | |
| template <class T, class OutputIterator>
 | |
| inline OutputIterator airy_bi_zero(
 | |
|                          int start_index,
 | |
|                          unsigned number_of_zeros,
 | |
|                          OutputIterator out_it)
 | |
| {
 | |
|    return airy_bi_zero<T>(start_index, number_of_zeros, out_it, policies::policy<>());
 | |
| }
 | |
| 
 | |
| }} // namespaces
 | |
| 
 | |
| #endif // BOOST_MATH_AIRY_HPP
 | 
