From laurence.thayer at navy.mil  Mon Oct 26 16:26:55 2009
From: laurence.thayer at navy.mil (larry)
Date: Mon Oct 26 16:26:58 2009
Subject: Problem with queue simulation
Message-ID: <60d2747d-c3a5-41be-8f8a-7f2a4395f2d6@m33g2000pri.googlegroups.com>

I am trying to update the single server queue model Gillett has in his
Intro to operations research book. I scrapped his use of Randu since
that has been discredited as random number generator, however this
seems to fill up too fast and I am getting values in the results for
queue size and time in system that seem excessive. Any ideas?

Option Explicit
Public Arr_rate As Single, Service_rate As Single, Clock As Single,
Arr_time As Single, Time_Sys As Single
Public Tot_Wait_Time As Single, Tot_Idle_time As Single, Ser_time As
Single, Time_Serv_fac As Single
Public Avg_Time_Sys As Single, Avg_Time_Fac As Single, Avg_Time_Q As
Single, Avg_Time_Wait As Single, Percent_Busy As Single
Public Max_cnt As Long, No_In_Q As Integer, Total_Q As Integer, No_arr
As Integer, Max_Q As Integer
Sub q_sim()
Arr_rate = Cells(2, 2).Value
Service_rate = Cells(3, 2).Value
Max_cnt = Cells(4, 2).Value
Clock = 0!
Arr_time = 0#
No_In_Q = 0
Total_Q = 0
Tot_Wait_Time = 0#
Tot_Idle_time = 0#
No_arr = 1
Max_Q = 0
Ser_time = Clock + (-1# / Service_rate) * Log(Rnd)

Do Until Clock > Max_cnt

Select Case Ser_time - Arr_time
    Case Is < 0
    If No_In_Q = 0 Then
        Clock = Ser_time
        Tot_Idle_time = Tot_Idle_time + (Arr_time - Clock)
        Clock = Arr_time
        No_arr = No_arr + 1
        Ser_time = Clock + (-1# / Service_rate) / Log(Rnd)
    Else
        Tot_Wait_Time = Tot_Wait_Time + No_In_Q * (Ser_time - Clock)
        Clock = Ser_time
        No_In_Q = No_In_Q - 1
        Ser_time = Clock + (-1# / Service_rate) / Log(Rnd)
    End If
    Case Is = 0
        Tot_Wait_Time = Tot_Wait_Time + No_In_Q * (Arr_time - Clock)
        Clock = Arr_time
        No_arr = No_arr + 1
        Arr_time = Clock + (-1# / Arr_rate) / Log(Rnd)
        Ser_time = Clock + (-1# / Service_rate) * Log(Rnd)
    Case Is > 0
    Tot_Wait_Time = Tot_Wait_Time + No_In_Q * (Arr_time - Clock)
        Clock = Arr_time
        No_arr = No_arr + 1
         No_In_Q = No_In_Q + 1
         If No_In_Q > Max_Q Then
            Max_Q = No_In_Q
         End If
            Total_Q = Total_Q + 1
            Arr_time = Clock + (-1# / Arr_rate) * Log(Rnd)
End Select
Loop

Time_Serv_fac = Clock - Tot_Idle_time
Cells(6, 2).Value = Time_Serv_fac
Time_Sys = Time_Serv_fac + Tot_Wait_Time
Cells(7, 2).Value = Time_Sys
Avg_Time_Sys = Time_Sys / No_arr
Cells(8, 2).Value = Avg_Time_Sys
Avg_Time_Fac = Time_Serv_fac / No_arr
Cells(9, 2).Value = Avg_Time_Fac
Avg_Time_Q = Tot_Wait_Time / No_arr
Cells(10, 2).Value = Avg_Time_Q
Avg_Time_Wait = Tot_Wait_Time / Total_Q
Cells(11, 2).Value = Avg_Time_Wait
Cells(12, 2).Value = Tot_Idle_time
Percent_Busy = Time_Serv_fac / Clock
Cells(13, 2).Value = Percent_Busy
Cells(14, 2).Value = Max_Q
End Sub

From jpwoodruff at gmail.com  Mon Oct 26 22:47:12 2009
From: jpwoodruff at gmail.com (jpwoodruff)
Date: Mon Oct 26 22:47:14 2009
Subject: what output for DEVS models?
Message-ID: <29f3cee7-2469-40ff-a33b-b7d71964b165@j9g2000prh.googlegroups.com>

Good Day all; I'm new here

I have just finished implementing the DEVS specification.
Now I'd like to do some simple example simulations, and I'm puzzled
what kind of output mechanisms to use.  I'm a hobbyist so my
experimental program will be *very* modest.

I will want to watch the flow of some events during a simulation
execution, and I choose to be limited to text only.  Clearly animated
graphics might be nice, but as a hobbyist I'm not the least interested
in building such a thing.

I made an atomic model called "editor" who can "print" a string that
arrives on its port.  But after there are 4 or 5 contributors, the
output gets pretty hard to watch.

I'd like a suggestion - how do simulators (if there are any) without
graphics communicate their results?

John

From umutdurak at gmail.com  Fri Oct 30 17:25:51 2009
From: umutdurak at gmail.com (Umut DURAK)
Date: Fri Oct 30 17:25:53 2009
Subject: JDMS Special Issue: Reusability,
	Interoperability and Composability in Air Warfare Simulations CfP
Message-ID: <73cc67b8-d395-4478-9014-c65fe8f75b00@g23g2000yqh.googlegroups.com>

Call For Papers
Journal of Defense Modeling and Simulation: Applications, Methodology,
Technology (JDMS)
Reusability, Interoperability and Composability in Air Warfare
Simulations
http://www.scs.org/PUBs/jdms/specialIssues/specialRICAWS.html

This special issue is devoted to publishing the results of recent
studies from air warfare simulation domain that have earned avant-
garde victories against the tough methodology problems of modeling and
simulation.

Publication Expected: Spring / Summer 2011

Guest Editors:
Dr. Umut DURAK
TUBITAK-SAGE
PK.16 06261 Mamak
Ankara, Turkey