VB.NET中的正态分布随机数答案

【问题标题】：Normal Distributed Random Number in VB.NETVB.NET中的正态分布随机数
【发布时间】：2013-10-08 06:08:43
【问题描述】：

有没有人知道如何在 vb.net 中制作正态分布的随机数？

谢谢

【问题讨论】：

【解决方案1】：

用法：

GaussNumDist(Mean, Standard Deviation, Sample Size)

下面的代码示例，它将用数字样本填充GaussNumArray()，其分布将具有指定的均值和标准差：

Imports System.Math

Module Module1
    Friend GaussNumArray() As Double
    Friend intICell As Long

    Friend Function GaussNumDist(ByVal Mean As Double, ByVal StdDev As Double, ByVal SampleSize As Integer)
        intICell = 1                'Loop variable

        ReDim GaussNumArray(SampleSize)

        Do While (intICell < (SampleSize + 1))
            Call NumDist(Mean, StdDev)
            Application.DoEvents()
        Loop
    End Function

    Sub NumDist(ByVal meanin As Double, ByVal sdin As Double)
        '---------------------------------------------------------------------------------
        'Converts uniform random numbers over the region 0 to 1 into Gaussian distributed
        'random numbers using Box-Muller algorithm.
        'Adapted from Numerical Recipes in C
        '---------------------------------------------------------------------------------

        'Defining variables
        Dim dblR1 As Double
        Dim dblR2 As Double
        Dim mean As Double
        Dim var As Double
        Dim circ As Double
        Dim trans As Double
        Dim dblY1 As Double
        Dim dblY2 As Double
        Dim Pi As Double
        Pi = 4 * Atan(1)

        'Get two random numbers
        dblR1 = (2 * UniformRandomNumber()) - 1
        dblR2 = (2 * UniformRandomNumber()) - 1

        circ = (dblR1 ^ 2) + (dblR2 ^ 2)        'Radius of circle

        If circ >= 1 Then       'If outside unit circle, then reject number
            Call NumDist(meanin, sdin)
            Exit Sub
        End If

        'Transform to Gaussian
        trans = Sqrt(-2 * Log(circ) / circ)

        dblY1 = (trans * dblR1 * sdin) + meanin
        dblY2 = (trans * dblR2 * sdin) + meanin

        GaussNumArray(intICell) = dblY1   'First number

        'Increase intICell for next random number
        intICell = (intICell + 1)

        GaussNumArray(intICell) = dblY2   'Second number

        'Increase intICell again ready for next call of ConvertNumberDistribution
        intICell = (intICell + 1)

    End Sub

    Friend Function UniformRandomNumber() As Double
        '-----------------------------------------------------------------------------------
        'Outputs random numbers with a period of > 2x10^18 in the range 0 to 1 (exclusive)
        'Implements a L'Ecuyer generator with Bays-Durham shuffle
        'Adapted from Numerical Recipes in C
        '-----------------------------------------------------------------------------------

        'Defining constants
        Const IM1 As Double = 2147483563
        Const IM2 As Double = 2147483399
        Const AM As Double = (1.0# / IM1)
        Const IMM1 As Double = (IM1 - 1.0#)
        Const IA1 As Double = 40014
        Const IA2 As Double = 40692
        Const IQ1 As Double = 53668
        Const IQ2 As Double = 52774
        Const IR1 As Double = 12211
        Const IR2 As Double = 3791
        Const NTAB As Double = 32
        Const NDIV As Double = (1.0# + IM1 / NTAB)
        Const ESP As Double = 0.00000012
        Const RNMX As Double = (1.0# - ESP)

        Dim iCell As Integer
        Dim idum As Double
        Dim j As Integer
        Dim k As Long
        Dim temp As Double

        Static idum2 As Long
        Static iy As Long
        Static iv(NTAB) As Long

        idum2 = 123456789
        iy = 0

        'Seed value required is a negative integer (idum)
        Randomize()
        idum = (-Rnd() * 1000)

        'For loop to generate a sequence of random numbers based on idum
        For iCell = 1 To 10
            'Initialize generator
            If (idum <= 0) Then
                'Prevent idum = 0
                If (-(idum) < 1) Then
                    idum = 1
                Else
                    idum = -(idum)
                End If
                idum2 = idum
                For j = (NTAB + 7) To 0
                    k = ((idum) / IQ1)
                    idum = ((IA1 * (idum - (k * IQ1))) - (k * IR1))
                    If (idum < 0) Then
                        idum = (idum + IM1)
                    End If
                    If (j < NTAB) Then
                        iv(j) = idum
                    End If
                Next j
                iy = iv(0)
            End If

            'Start here when not initializing
            k = (idum / IQ1)
            idum = ((IA1 * (idum - (k * IQ1))) - (k * IR1))
            If (idum < 0) Then
                idum = (idum + IM1)
            End If
            k = (idum2 / IQ2)
            idum2 = ((IA2 * (idum2 - (k * IQ2))) - (k * IR2))
            If (idum2 < 0) Then
                idum2 = idum2 + IM2
            End If
            j = (iy / NDIV)
            iy = (iv(j) - idum2)
            iv(j) = idum
            If (iy < 1) Then
                iy = (iy + IMM1)
            End If
            temp = AM * iy
            If (temp <= RNMX) Then
                'Return the value of the random number
                UniformRandomNumber = temp
            End If
        Next iCell
    End Function
End Module

【讨论】：

我已经尝试过您在上面分享的代码。它工作正常，我可以制作具有自定义大小、平均值和标准偏差的随机数。然后我向谷歌询问有关正态分布随机数的另一个参考，并找到了这个链接 [链接] (codeproject.com/Articles/15102/…)。不幸的是，我分享的链接上的代码是用 C# 制作的。有没有机会在我分享给我的程序的链接上使用正态分布代码，它是用 VB.NET 先生 @Roger Rowland 编写的？
@mano - 你可以convert it here，如果我的回答有帮助，也许你可以通过点击复选标记将其变为绿色和/或投票来接受它。

【解决方案2】：

您可以使用以下行

Dim x1 as Double = MathNet.Numerics.Distributions.Normal.Sample(MEAN, STDEV)

Math.Net Numeric 包可以使用以下 NuGet 命令安装

Install-Package MathNet.Numerics -Version 4.9.0

您可以在NuGet site找到更多信息

【讨论】：