Supply Chain Trading Analysis and Instrumentation Toolkit

AIT Shortcuts

Introduction

AIT is a collection of tools useful for analyzing historical log files from 2005 TAC SCM servers. The tools provided allow in-depth analysis of B2B and B2C interactions through an easy to read graphical front end. The current distribution includes a GUI for the tools, and the tools themselves which utilize the included library, JFreeChart (this library is freely available under the LGPL and the source code can be accessed at the JFreeChart website). Our tool also uses a binary version of the SICS TAC SCM Game Data Toolkit in our distribution, this library is provided in our distribution, and the full source code and the conditions for use are available to you at http://www.sics.se/tac/.

The toolbox and its front-end were designed and written by the e-Supply Chain Management Lab's CMieux team lead by Michael Benisch and Norman Sadeh, with several undergraduate students.

About the Instrumentation Tools

The tools currently provide facilities for:

• Automatically downloading and parsing game files from any 2005 (or earlier) TAC SCM Server.
• Focusing analysis on any agent in the game by highlighting its performance.
• Graphing agent reputation for all suppliers over the entire game.
• Displaying agent lead times, prices, and quantities graphically for all supplier orders.
• Graphing average daily component prices of all agents over the entire game by component type.
• Displaying a market share pie-chart for any combination of the different customer markets captured by each agent.
• Changing the window in which market share is reported from any start point, to any end point in the game.
• Graphing average daily winning and losing bid prices of all agents over the entire game by product grade (high, medium, and low).
• Choosing between displaying any prices as fractions of component base prices (a relative measure of quality), or dollars for different kinds of analysis.

The software is all implemented in Java 1.5 using the Swing GUI development API and JFreeChart. Consequently, this software should run accross platforms, but we have not as of yet tested beyond Linux, Windows XP, and Mac OSX. Presently we will release only the binaries of our code, but we hope to augment this with a full source release soon.

back to top

Citation

Benisch, et. al., "CMieux Analysis and Instrumentation Toolkit
for TAC SCM", Carnegie Mellon University School of Computer Science
Technical Report CMU-ISRI-05-127, 2005.

Here is a BiBTeX entry:

   @techreport{ benisch05ait,
	author = "Michael Benisch and James Andrews and David
	Bangerter and Timothy Kirchner and Benjamin Tsai and Norman
	Sadeh",
	title = "CMieux Analysis and Instrumentation Toolkit for TAC SCM",
	number = "CMU-ISRI-05-127",
	institution = "School of Computer Science, Carnegie Mellon University",
	month = "September",
	year = 2005
    }

back to top

Obtaining the Tools

A Zip file containing the binaries and libraries for the tools can be downloaded here. After downloading the file, unzip it and you will find the following contents:

• instrumentation.jar is an executable jar file that contains all of the binaries for the instrumentation tools.
• lib/sicslib.jar is jar file containing the libraries for log file parser implementation developped by SICS.
• lib/[jfreechart|jcommon]....jar are jar files containing the libraries for JFreeChart.
• default.conf is a text file that specifies the default settings of the GUI when loaded to ease repeated analysis.
• README.txt is a readme file with quick start and copyright information.

back to top

Quick Start: Using the Software

Invoke the GUI using the following command from the directory containing the contents of the Zip file:

[ait_direcory]% java -jar instrumentation.jar

• This will bring up the main UI window:
  
  
  
  
• Once the GUI is invoked specify the game number, server location, and agent to focus on in the specified fields:
  
  
  
  
• Press the "Gather Data" button on the GUI to begin downloading and parsing the specified game data. The game data will be downloaded from the server and stored locally at the path:

  [ait_directory]/games/[server_name]/[game_number]/game.slg.gz

  (this only be done if the game data does not already exist locally). The log file will then be parsed and all information needed for analysis will be gathered.

• Choose one of the analysis tools from the B2B or B2C groups by pressing the corresponding button.
• For more comprehensive documentation please see the PDF documentation found on this site.

We cannot garauntee that this software is bug free and entirely correct. Feel free to send email asking for help to Michael Benisch; bug reports are appreciated.

back to top