Random.Sample 方法
定義
重要
部分資訊涉及發行前產品，在發行之前可能會有大幅修改。 Microsoft 對此處提供的資訊，不做任何明確或隱含的瑕疵擔保。
回傳一個介於 0.0 到 1.0 之間的隨機浮點數。
protected:
virtual double Sample();protected virtual double Sample();abstract member Sample : unit -> double
override this.Sample : unit -> doubleProtected Overridable Function Sample () As Double傳回
一個大於或等於0.0且小於1.0的雙精度浮點數。
範例
以下範例從 衍生出一個類別 Random ，並覆寫該 Sample 方法以產生隨機數分布。 此分布與基 Sample 類方法產生的均勻分布不同。
using System;
// This derived class converts the uniformly distributed random
// numbers generated by base.Sample() to another distribution.
public class RandomProportional : Random
{
// The Sample method generates a distribution proportional to the value
// of the random numbers, in the range [0.0, 1.0].
protected override double Sample()
{
return Math.Sqrt(base.Sample());
}
public override int Next()
{
return (int) (Sample() * int.MaxValue);
}
}
public class RandomSampleDemo
{
static void Main()
{
const int rows = 4, cols = 6;
const int runCount = 1000000;
const int distGroupCount = 10;
const double intGroupSize =
((double)int.MaxValue + 1.0) / (double)distGroupCount;
RandomProportional randObj = new RandomProportional();
int[ ] intCounts = new int[ distGroupCount ];
int[ ] realCounts = new int[ distGroupCount ];
Console.WriteLine(
"\nThe derived RandomProportional class overrides " +
"the Sample method to \ngenerate random numbers " +
"in the range [0.0, 1.0]. The distribution \nof " +
"the numbers is proportional to their numeric values. " +
"For example, \nnumbers are generated in the " +
"vicinity of 0.75 with three times the \n" +
"probability of those generated near 0.25.");
Console.WriteLine(
"\nRandom doubles generated with the NextDouble() " +
"method:\n");
// Generate and display [rows * cols] random doubles.
for (int i = 0; i < rows; i++)
{
for (int j = 0; j < cols; j++)
Console.Write("{0,12:F8}", randObj.NextDouble());
Console.WriteLine();
}
Console.WriteLine(
"\nRandom integers generated with the Next() " +
"method:\n");
// Generate and display [rows * cols] random integers.
for (int i = 0; i < rows; i++)
{
for (int j = 0; j < cols; j++)
Console.Write("{0,12}", randObj.Next());
Console.WriteLine();
}
Console.WriteLine(
"\nTo demonstrate the proportional distribution, " +
"{0:N0} random \nintegers and doubles are grouped " +
"into {1} equal value ranges. This \n" +
"is the count of values in each range:\n",
runCount, distGroupCount);
Console.WriteLine(
"{0,21}{1,10}{2,20}{3,10}", "Integer Range",
"Count", "Double Range", "Count");
Console.WriteLine(
"{0,21}{1,10}{2,20}{3,10}", "-------------",
"-----", "------------", "-----");
// Generate random integers and doubles, and then count
// them by group.
for (int i = 0; i < runCount; i++)
{
intCounts[ (int)((double)randObj.Next() /
intGroupSize) ]++;
realCounts[ (int)(randObj.NextDouble() *
(double)distGroupCount) ]++;
}
// Display the count of each group.
for (int i = 0; i < distGroupCount; i++)
Console.WriteLine(
"{0,10}-{1,10}{2,10:N0}{3,12:N5}-{4,7:N5}{5,10:N0}",
(int)((double)i * intGroupSize),
(int)((double)(i + 1) * intGroupSize - 1.0),
intCounts[ i ],
((double)i) / (double)distGroupCount,
((double)(i + 1)) / (double)distGroupCount,
realCounts[ i ]);
}
}
/*
This example of Random.Sample() displays output similar to the following:
The derived RandomProportional class overrides the Sample method to
generate random numbers in the range [0.0, 1.0). The distribution
of the numbers is proportional to the number values. For example,
numbers are generated in the vicinity of 0.75 with three times the
probability of those generated near 0.25.
Random doubles generated with the NextDouble() method:
0.59455719 0.17589882 0.83134398 0.35795862 0.91467727 0.54022658
0.93716947 0.54817519 0.94685080 0.93705478 0.18582318 0.71272428
0.77708682 0.95386216 0.70412393 0.86099417 0.08275804 0.79108316
0.71019941 0.84205103 0.41685082 0.58186880 0.89492302 0.73067715
Random integers generated with the Next() method:
1570755704 1279192549 1747627711 1705700211 1372759203 1849655615
2046235980 1210843924 1554274149 1307936697 1480207570 1057595022
337854215 844109928 2028310798 1386669369 2073517658 1291729809
1537248240 1454198019 1934863511 1640004334 2032620207 534654791
To demonstrate the proportional distribution, 1,000,000 random
integers and doubles are grouped into 10 equal value ranges. This
is the count of values in each range:
Integer Range Count Double Range Count
------------- ----- ------------ -----
0- 214748363 10,079 0.00000-0.10000 10,148
214748364- 429496728 29,835 0.10000-0.20000 29,849
429496729- 644245093 49,753 0.20000-0.30000 49,948
644245094- 858993458 70,325 0.30000-0.40000 69,656
858993459-1073741823 89,906 0.40000-0.50000 90,337
1073741824-1288490187 109,868 0.50000-0.60000 110,225
1288490188-1503238552 130,388 0.60000-0.70000 129,986
1503238553-1717986917 149,231 0.70000-0.80000 150,428
1717986918-1932735282 170,234 0.80000-0.90000 169,610
1932735283-2147483647 190,381 0.90000-1.00000 189,813
*/
open System
// This derived class converts the uniformly distributed random
// numbers generated by base.Sample() to another distribution.
type RandomProportional() =
inherit Random()
// The Sample method generates a distribution proportional to the value
// of the random numbers, in the range [0.0, 1.0].
override _.Sample() =
sqrt (base.Sample())
override this.Next() =
this.Sample() * float Int32.MaxValue
|> int
let [<Literal>] rows = 4
let [<Literal>] cols = 6
let [<Literal>] runCount = 1000000
let [<Literal>] distGroupCount = 10
let intGroupSize =
(float Int32.MaxValue + 1.0) / float distGroupCount
let randObj = RandomProportional()
printfn """
The derived RandomProportional class overrides the Sample method to
generate random numbers in the range [0.0, 1.0]. The distribution
of the numbers is proportional to their numeric values. For example,
numbers are generated in the vicinity of 0.75 with three times the
probability of those generated near 0.25."""
printfn "\nRandom doubles generated with the NextDouble() method:\n"
// Generate and display [rows * cols] random doubles.
for _ = 1 to rows do
for _ = 1 to cols do
printf $"{randObj.NextDouble(),12:F8}"
printfn ""
printfn "\nRandom integers generated with the Next() method:\n"
// Generate and display [rows * cols] random integers.
for _ = 1 to rows do
for _ = 1 to cols do
printf $"{randObj.Next(),12}"
printfn ""
printfn $"""
To demonstrate the proportional distribution, {runCount:N0} random
integers and doubles are grouped into {distGroupCount} equal value ranges. This
is the count of values in each range:
"""
printfn $"""{"Integer Range",21}{"Count",10}{"Double Range",20}{"Count",10}"""
printfn $"""{"-------------",21}{"-----",10}{"------------",20}{"-----",10}"""
// Generate random integers and doubles, and then count them by group.
let intCounts =
Array.init runCount (fun _ ->
(randObj.Next() |> float) / float intGroupSize
|> int )
|> Array.countBy id
|> Array.map snd
let realCounts =
Array.init runCount (fun _ ->
randObj.NextDouble() * float distGroupCount
|> int )
|> Array.countBy id
|> Array.map snd
// Display the count of each group.
for i = 0 to distGroupCount - 1 do
Console.WriteLine(
"{0,10}-{1,10}{2,10:N0}{3,12:N5}-{4,7:N5}{5,10:N0}",
int(float i * intGroupSize),
int(float (i + 1) * intGroupSize - 1.0),
intCounts.[i],
(float i) / float distGroupCount,
float (i + 1) / float distGroupCount,
realCounts.[i])
(*
This example of Random.Sample() displays output similar to the following:
The derived RandomProportional class overrides the Sample method to
generate random numbers in the range [0.0, 1.0). The distribution
of the numbers is proportional to the number values. For example,
numbers are generated in the vicinity of 0.75 with three times the
probability of those generated near 0.25.
Random doubles generated with the NextDouble() method:
0.59455719 0.17589882 0.83134398 0.35795862 0.91467727 0.54022658
0.93716947 0.54817519 0.94685080 0.93705478 0.18582318 0.71272428
0.77708682 0.95386216 0.70412393 0.86099417 0.08275804 0.79108316
0.71019941 0.84205103 0.41685082 0.58186880 0.89492302 0.73067715
Random integers generated with the Next() method:
1570755704 1279192549 1747627711 1705700211 1372759203 1849655615
2046235980 1210843924 1554274149 1307936697 1480207570 1057595022
337854215 844109928 2028310798 1386669369 2073517658 1291729809
1537248240 1454198019 1934863511 1640004334 2032620207 534654791
To demonstrate the proportional distribution, 1,000,000 random
integers and doubles are grouped into 10 equal value ranges. This
is the count of values in each range:
Integer Range Count Double Range Count
------------- ----- ------------ -----
0- 214748363 10,079 0.00000-0.10000 10,148
214748364- 429496728 29,835 0.10000-0.20000 29,849
429496729- 644245093 49,753 0.20000-0.30000 49,948
644245094- 858993458 70,325 0.30000-0.40000 69,656
858993459-1073741823 89,906 0.40000-0.50000 90,337
1073741824-1288490187 109,868 0.50000-0.60000 110,225
1288490188-1503238552 130,388 0.60000-0.70000 129,986
1503238553-1717986917 149,231 0.70000-0.80000 150,428
1717986918-1932735282 170,234 0.80000-0.90000 169,610
1932735283-2147483647 190,381 0.90000-1.00000 189,813
*)
' This derived class converts the uniformly distributed random
' numbers generated by base.Sample() to another distribution.
Public Class RandomProportional
Inherits Random
' The Sample method generates a distribution proportional to the value
' of the random numbers, in the range [0.0, 1.0].
Protected Overrides Function Sample() As Double
Return Math.Sqrt(MyBase.Sample())
End Function
Public Overrides Function [Next]() As Integer
Return Sample() * Integer.MaxValue
End Function
End Class
Module RandomSampleDemo
Sub Main()
Const rows As Integer = 4, cols As Integer = 6
Const runCount As Integer = 1000000
Const distGroupCount As Integer = 10
Const intGroupSize As Double = _
(CDbl(Integer.MaxValue) + 1.0) / _
CDbl(distGroupCount)
Dim randObj As New RandomProportional()
Dim intCounts(distGroupCount) As Integer
Dim realCounts(distGroupCount) As Integer
Dim i As Integer, j As Integer
Console.WriteLine(vbCrLf & _
"The derived RandomProportional class overrides " & _
"the Sample method to " & vbCrLf & _
"generate random numbers in the range " & _
"[0.0, 1.0]. The distribution " & vbCrLf & _
"of the numbers is proportional to their numeric " & _
"values. For example, " & vbCrLf & _
"numbers are generated in the vicinity of 0.75 " & _
"with three times " & vbCrLf & "the " & _
"probability of those generated near 0.25.")
Console.WriteLine(vbCrLf & _
"Random doubles generated with the NextDouble() " & _
"method:" & vbCrLf)
' Generate and display [rows * cols] random doubles.
For i = 0 To rows - 1
For j = 0 To cols - 1
Console.Write("{0,12:F8}", randObj.NextDouble())
Next j
Console.WriteLine()
Next i
Console.WriteLine(vbCrLf & _
"Random integers generated with the Next() " & _
"method:" & vbCrLf)
' Generate and display [rows * cols] random integers.
For i = 0 To rows - 1
For j = 0 To cols - 1
Console.Write("{0,12}", randObj.Next())
Next j
Console.WriteLine()
Next i
Console.WriteLine(vbCrLf & _
"To demonstrate the proportional distribution, " & _
"{0:N0} random " & vbCrLf & _
"integers and doubles are grouped into {1} " & _
"equal value ranges. This " & vbCrLf & _
"is the count of values in each range:" & vbCrLf, _
runCount, distGroupCount)
Console.WriteLine("{0,21}{1,10}{2,20}{3,10}", _
"Integer Range", "Count", "Double Range", "Count")
Console.WriteLine("{0,21}{1,10}{2,20}{3,10}", _
"-------------", "-----", "------------", "-----")
' Generate random integers and doubles, and then count
' them by group.
For i = 0 To runCount - 1
intCounts(Fix(CDbl(randObj.Next()) / _
intGroupSize)) += 1
realCounts(Fix(randObj.NextDouble() * _
CDbl(distGroupCount))) += 1
Next i
' Display the count of each group.
For i = 0 To distGroupCount - 1
Console.WriteLine( _
"{0,10}-{1,10}{2,10:N0}{3,12:N5}-{4,7:N5}{5,10:N0}", _
Fix(CDbl(i) * intGroupSize), _
Fix(CDbl(i + 1) * intGroupSize - 1.0), _
intCounts(i), _
CDbl(i) / CDbl(distGroupCount), _
CDbl(i + 1) / CDbl(distGroupCount), _
realCounts(i))
Next i
End Sub
End Module
' This example of Random.Sample() generates output similar to the following:
'
' The derived RandomProportional class overrides the Sample method to
' generate random numbers in the range [0.0, 1.0]. The distribution
' of the numbers is proportional to their numeric values. For example,
' numbers are generated in the vicinity of 0.75 with three times
' the probability of those generated near 0.25.
'
' Random doubles generated with the NextDouble() method:
'
' 0.28377004 0.75920598 0.33430371 0.66720626 0.97080243 0.27353772
' 0.17787962 0.54618410 0.08145080 0.56286100 0.99002910 0.64898614
' 0.27673277 0.99455281 0.93778966 0.76162002 0.70533771 0.44375798
' 0.55939883 0.87383136 0.66465779 0.77392566 0.42393411 0.82409159
'
' Random integers generated with the Next() method:
'
' 1364479914 1230312341 1657373812 1526222928 988564704 700078020
' 1801013705 1541517421 1146312560 338318389 1558995993 2027260859
' 884520932 1320070465 570200106 1027684711 943035246 2088689333
' 630809089 1705728475 2140787648 2097858166 1863010875 1386804198
'
' To demonstrate the proportional distribution, 1,000,000 random
' integers and doubles are grouped into 10 equal value ranges. This
' is the count of values in each range:
'
' Integer Range Count Double Range Count
' ------------- ----- ------------ -----
' 0- 214748363 9,892 0.00000-0.10000 9,928
' 214748364- 429496728 30,341 0.10000-0.20000 30,101
' 429496729- 644245093 49,958 0.20000-0.30000 49,964
' 644245094- 858993458 70,099 0.30000-0.40000 70,213
' 858993459-1073741823 90,801 0.40000-0.50000 89,553
' 1073741824-1288490187 109,699 0.50000-0.60000 109,427
' 1288490188-1503238552 129,438 0.60000-0.70000 130,339
' 1503238553-1717986917 149,886 0.70000-0.80000 150,000
' 1717986918-1932735282 170,338 0.80000-0.90000 170,128
' 1932735283-2147483647 189,548 0.90000-1.00000 190,347
備註
若要產生不同的隨機分布或不同的隨機數產生原理，請從該Random類別推導出類別並覆蓋該方法。Sample
Important
該 Sample 方法為 protected，表示它僅能在本類別及其衍生類別中 Random 存取。 要從實 Random 例產生介於 0 到 1 之間的隨機數，請呼叫該 NextDouble 方法。
給繼承者的注意事項
從 .NET Framework 2.0 版本開始，若您從 Random 衍生出類別並覆寫 Sample() 方法，則 Sample() 方法的導出類別實作所提供的分布不會用於呼叫以下方法的基底類別實作：
Next() 方法。
若 （Next(Int32, Int32)
maxValue- ） 大於minValue，則該方法。
取而代之的是基底 Random 類別所提供的均勻分布。 這種行為能提升班級整體 Random 表現。 若要修改此行為以呼叫導出類別中方法的實作 Sample() ，你也必須覆蓋這三個成員的行為。 此範例提供圖例。
