Ideas for Simple A.I. Builds

[Pete]

So part of A.I. is getting a computer to exhibit critical thinking.

I guess we could start off with some rudimentary approaches, for example:

Code: (Select All)

Randomize Timer

Palette 7, 63

Color 0, 7: Cls

View Print 1 To 23

Dim As Integer red, green, blue, mred, mgreen, mblue, i, cnt, pass(3)

score = 100

Do

    cnt = cnt + 1

    red = Rnd * 10 - 4

    green = Rnd * 10 - 6

    blue = Rnd * 10 - 8

    i = Int(Rnd * 3) + 1

    Print i;

    If pass(i) Then action$ = "Pass" Else action$ = "Catch"

    Select Case i

        Case 1

            Color 4, 7: Print "RED "; red; action$

            If action$ = "Catch" Then

                score = score + red

                mred = mred + red

            End If

        Case 2

            Color 2, 7: Print "GREEN "; green; action$

            If action$ = "Catch" Then

                score = score + green

                mgreen = mgreen + green

            End If

        Case 3

            Color 1, 7: Print "BLUE "; blue; action$

            If action$ = "Catch" Then

                score = score + blue

                mblue = mblue + blue

            End If

    End Select

    Color 0, 7

    x = CsrLin

    Locate 25, 1: Print "Score = "; LTrim$(Str$(score)); "    ";

    Locate x, 1



    If cnt = 10 Then

        If mred < mblue And mred < mgreen Then If pass(2) + pass(3) <> 5 Then pass(1) = 1 ' Don't catch red apples.

    ElseIf mgreen < mred And mgreen < mblue Then If pass(1) + pass(3) <> 4 Then pass(2) = 2 ' Don't catch green apples.

    ElseIf mblue < mred And mblue < mgreen Then If pass(1) + pass(2) <> 3 Then pass(3) = 3 ' Don't catch blue apples.

        cnt = 0

    End If

    Sleep

    If InKey$ = Chr$(13) Then Run

Loop

The computer memorizes and evaluates values, and decides to only catch red apples, fairly soon.

Refinements would include ways to deduce this faster, and to widen the sampling if we gave it feedback using greater variance with its results vs the best possible outcome. I mean real A.I. means the computer would keep refining its calculation method until it reached the most efficient method.

Anyway, just having some fun and wondering if anyone else has made anything along A.I. lines they would like to share, or just discuss in this thread.

I'll be back a bit later today to catch up!

Pete