箱装优化最佳配合

时间:2013-03-26 11:48:00

标签: algorithm optimization cut bin-packing

我有这个Bin Packing代码:

Public Class Cortar
    Private _Cortes() As Integer
    Public BarrasCortes()() As Integer
    Private _TamanhoBarra As Integer = 100

    Public Property TamanhoBarra() As Integer
        Get
            Return _TamanhoBarra
        End Get
        Set(ByVal Value As Integer)
            _TamanhoBarra = Value
        End Set
    End Property

    Public Property Cortes() As Integer()
        Get
            Return _Cortes
        End Get
        Set(ByVal Value As Integer())
            _Cortes = Value
        End Set
    End Property

    Public Sub MelhorCorte()
        'Checks to make sure everything is initialized
        If Cortes Is Nothing Then Exit Sub

        Dim CortesCopy(Cortes.GetUpperBound(0)) As Integer
        ReDim BarrasCortes(0)
        'Bin Number we are on, Bin Element we are on, Amount placed in the current Bin
        Dim BinNumber, BinElement, BinCount As Integer
        Dim BestBin, BestBinAmount As Integer
        Dim i, j, k As Integer

        'Make a copy of the array incase we need to sort it
        DeepCopyArray(Cortes, CortesCopy)

        'Declare the first element in the first Bin
        ReDim BarrasCortes(0)(0)

        'Loop through each Element and place in a Bin
        For i = 0 To CortesCopy.GetUpperBound(0)
            BestBin = -1
            BestBinAmount = -1

            For j = 0 To BinNumber
                BinElement = BarrasCortes(j).GetUpperBound(0)

                'Count the amount placed in this Bin
                BinCount = 0
                For k = 0 To BinElement
                    BinCount += BarrasCortes(j)(k)
                Next

                'Find the most full Bin that can hold this Element
                If BestBinAmount < BinCount AndAlso BinCount + CortesCopy(i) <= Me.TamanhoBarra Then
                    BestBinAmount = BinCount
                    BestBin = j
                End If
            Next


            If BestBin = -1 Then
                'There wasn't room for the Element in any existing Bin
                'Create a new Bin
                ReDim Preserve BarrasCortes(BinNumber + 1)
                BinNumber += 1

                'Initialize first element of new bin
                ReDim BarrasCortes(BinNumber)(1)
                BinElement = 0

                BarrasCortes(BinNumber)(BinElement) = CortesCopy(i)
            Else
                'There's room for this Element in an existing Bin
                'Place Element in "Best Bin"
                BinElement = BarrasCortes(BestBin).GetUpperBound(0)
                ReDim Preserve BarrasCortes(BestBin)(BinElement + 1)
                BarrasCortes(BestBin)(BinElement) = CortesCopy(i)
            End If
        Next

        'All Cortes have been place, now we go back and remove unused Cortes
        For i = 0 To BinNumber
            For j = 0 To BarrasCortes(i).GetUpperBound(0)
                If BarrasCortes(i)(j) = 0 Then
                    ReDim Preserve BarrasCortes(i)(j - 1)
                End If
            Next
        Next

        GC.Collect()
    End Sub

    Private Sub DeepCopyArray(ByVal ArrayStart() As Integer, ByVal ArrayEnd() As Integer)
        Dim i As Integer

        For i = 0 To ArrayStart.GetUpperBound(0)
            ArrayEnd(i) = ArrayStart(i)
        Next
    End Sub
End Class

这很好,但有问题:

如果我在C#中定义:

var obj= new Cortar();
obj.TamanhoBarra = 100;
obj.Cortes = new int[] {48, 48, 26, 26};
obj.MelhorCorte();
int[][] x = obj.BarrasCortes;

结果是:

x[0] = {48,48} //rest 4 to 100 (TamanhoBarra)
x[1] = {26,26} //rest 48 to 100 (TamanhoBarra)

但是,如果我只是将数组中元素的顺序更改为

obj.Cortes = new int[] {48, 26, 48, 26};

结果将是:

x[0] = {48,26,26} //rest 0 to 100 (TamanhoBarra) = best optimization
x[1] = {48} //rest 52 to 100 (TamanhoBarra)

问题是:

如何最佳选择吧?

最佳的不对称是直到最大尺寸的元素之和,而休息时间较少。

在上述情况下很容易。

但如果我有:

obj.Cortes = new int[] {48, 48, 48, 48, 48, 48, 48, 48, 48, 48, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26 };

结果将是:

x[0] = {48,48} //rest 4
x[1] = {48,48} //rest 4
x[2] = {48,48} //rest 4
x[3] = {48,48} //rest 4
x[4] = {48,48} //rest 4
x[5] = {26,26,26} //rest 22
x[6] = {26,26,26} //rest 22
x[7] = {26,26,26} //rest 22
x[8] = {26} //rest 74

很多酒吧和全部损失

但如果我像这样排列数组:

obj.Cortes = new int[] {48, 26, 26, 48, 26, 26, 48, 26, 26, 48, 26, 26, 48, 26, 26, 48, 48, 48, 48, 48 };

结果是:

x[0] = {48,26,26} //rest 0
x[1] = {48,26,26} //rest 0
x[2] = {48,26,26} //rest 0
x[3] = {48,26,26} //rest 0
x[4] = {48,26,26} //rest 0
x[5] = {48,48} //rest 4
x[6] = {48,48} //rest 4
x[7] = {48} //rest 52

这是最好的解决方案!

有什么想法吗?

1 个答案:

答案 0 :(得分:0)

只需少量商品即可尝试。然后你可以尝试动态编程。