Friday, February 27, 2009
I have to give props to my alma-mater Virginia Tech and the department from which I received my Industrial Engineering degree. The Virginia Tech Grado Department of Industrial and Systems Engineering was awarded a $40,000 grant from UPS. This gives them the 13th consecutive year of grants awarded to Virginia Tech from UPS.
I was a student at Virginia Tech while Dr. Casali was the department head. It was a great time to be part of the ISE department as it was growing. Dr. Casali was a large influence on the success of the ISE department. Contratulations to Dr. Casali and the ISE Team at Virginia Tech.
Wednesday, February 25, 2009
I make no secret that I am Open Source advocate. I blog about Open Source software in the Industrial Engineering and Operations Research realm. I also post comments on other blogs in both Open Source and Operations Research about software tools that are free to use and distribute. My reason for doing this is very simple. I want to make others aware of the inherit freedoms and innovation that Open Source software brings to the Industrial Engineering and Operations Research community. I believe there is much to discover and so little available.
There seems to be the haves and have nots in the Operations Research world currently. First, those that have the proprietary software tools to do incredible decision and predictive analysis. Second, those that rely on their education, intuition, and experience with ubiquitous or free software tools to complete the job. Most of my career has fallen in the latter as you can guess. I have made the most of what I was given to complete my work. I believe it is good work yet I know a lot of times it could be even better if I had the right tools.
I'm not bitter that organizations exist to sell their products. I believe there is nothing wrong with that philosophy. Yet I also believe that holding software in a proprietary realm stiffles innovation and advancement. To me software is just a set of rules, instructions, and algorithms. Software is akin to the Pythagorean Theorem or Little's Law. Where would we be today if Sir Isaac Newton had copyrighted calculus? Just as the Reinassance Era was a boom to modern thinking I believe the Free Libre and Open Source Software era can be a boom to new innovative ideas.
A lot of people are debating the merits of this today. There is an excellent example this week with Amazon's release of Kindle 2. Tim O'Reilly, of O'Reilly publishing fame, debates the merits of closing the source to the Kindle 2 and feels it should be opened. Yet Matt Asay at CNET feels Amazon should stick to their guns of being a content provider and push forward with being a proprietary system. They both have merit. To me opening the Kindle can still give Amazon its value as a content provider but also leverage the community to help generate new innovative ideas. This is a very interesting debate on the basis of core competecies versus innovation to new market trends. Only time will tell but to me the e-book will more than likely go the way of the World Wide Web as Tim O'Reilly pointed out.
Jim Orlin has a really interesting post on what Operations Research professionals should know and how will they learn it? Jim brings up good points about being subject matter experts in all knowledgable areas of Operations Research. A lot of times the Operations Research analyst is going to have to learn on the job. How are we to provide those resources and tools that they need on a continual basis?
Again only time will tell if the Operations Research realm is hindered by not opening software. I believe we are on the precipice of great things to come. The current recession has brought new thinking and demand to decision analysis. There is definitely greater appreciation for us "statisticians".
Monday, February 23, 2009
The Industrial Engineer often needs to explain a process flow, project diagram, or simple layouts. No one can replace the simple pencil and paper sketches but the computer has bridged the gap from an idea concept to professional presentation.
There are many Open Source computer aided diagramming tools available for the Industrial Engineer. Here is a list of diagramming tools that you can add to your tool bag free of cost.
Dia - excellent drawing program with strong community support. Available on Linux and Windows.
Kivio - the KOffice version of a drawing, flowcharting, and diagramming application.
OpenOffice Draw - integrated with the OpenOffice software suite. This diagramming tool is very easy to use.
There are even browser based, server-side applications that you can find on the internet. All you need is a computer with an internet connection and a decent web browser such as Firefox, Opera, or Chrome. These web based diagramming tools include the following.
Wednesday, February 18, 2009
While looking for open source alternatives I found this great website that runs a query to just about any common commercial software. The website is osalt.com. You may select from many categories of software applications. You can also select a specific commercial application and it will list the open source alternatives underneath it.
This just goes to show that there are many options available in the open source community. Maybe we can start add Operations Research software applications to this website.
Monday, February 16, 2009
There is a great post on Intechne Blog called Living in Interesting Times. The post is a look back on Operations Research software changes within the industry and the organizational movement. It really is a nice change of pace to see so many changes in the news.
There has been a lot of activity recently with concern to decision and optimization software. Organizations are trying to leverage different ways to expand their sphere of influence. I believe there is a trend in software services from these organizations. Here are my thoughts on the future of optimization software in Operations Research.
1. Data Mining and Optimization Modeling Software Convergence. There seems to be a trend to package optimization software applications and software frameworks with current data mining applications. Microsoft seems to be the best example of this with the Microsoft Solver Foundation. Open Source software can easily be applied to this idea as all of the code is available for modifying and redistributing in usable plugins or API's. Examples include OpenOpt and GLPK. SAS is also working to upgrade their Optimization team. I will not be surprised if IBM follows suit, considering their acquisition of ILOG, with a CPLEX API in the future. Enabling ubiquitous software applications like Microsoft Office or OpenOffice with optimization software will leverage users with new, low cost and varied ways to drive business decisions.
2. Decision Analysis will trump Data Analysis. Providing data warehouse services will not be enough for most customers of these software products and services. Organizations that use these software applications will want more robust ways to sift through the data and provide information. The optimization modeling components of rules engines and business rules will be heavily relied upon to get the most of the given data.
You can already see this happening today. In the ILOG Events Blog they have already talked about IBM/ILOG strategies during DIALOG. From the blog article...
...IBM is committed to all the products and to them as standalone products but also looking to integrate into new and existing IBM products.
...(IBM's survey of CEO's) 57% more data next year, average employees spend 2 hours looking for data. How to turn data into information and turn the insight into action.
Thursday, February 12, 2009
Since I am learning more about business rules I thought I would share some resources from around the internet. There is definitely a lot of information on the subject. It does take some effort to sift between the advertising and the good content.
IBM developerWorks has a great article by Arun Chhatpar on an Introduction to Business Rules. Incidentatly IBM developerWorks has a great wealth of information on Open Source software. Additionally here is the Wikipedia entry on Business Rules. The Business Rules Group, a non-commercial IT peer group, has some good examples of Business Rules.
Please share other sources of Rules Engines and Business Rules.
Monday, February 9, 2009
I've seen much talk on the internet about advancement of Rules Engines and Business Rules Management Systems (BRMS). As you can imagine there are Open Source options for Rules Engines and business rules systems. It seems the top players in Rules Engines are written mostly in the Java programming language which fairly recently was made Open Source by Sun Microsystems. Yet with Open Source software the possibilties could be endless with the tools used to develop Rules Engines.
Some of the top Open Source players in the Rules Engine business include:
The interesting thing about Rules Engines and Business Rules Management Systems is they most invariably use a rules repository. This repository could be setup on a client/server network which is very easy to setup with the Open Source Linux operating system.
For another resource, Business Review Online's Open Source Blog had a review last year of the Top 10 Best Open Source Rules Engines.
Now I have to admit that I am fairly new to BRMS and Rules Engines. I tend to deal more with the modelling and decision support systems of the Rules Engines. I would love to learn more about this area.
Thursday, February 5, 2009
One of the great traits of developing software in the Open Source community is that you can build on the successes of others. Luiz Bettoni has done just that with his linear programming tool known as GUSEK. Recently Luiz released version 0.2.2 of GUSEK to the community. GUSEK is a linear programming modeling interface to the GLPK. GUSEK is freely availabe to download. It is designed for the a Windows operating system. IEOR Tools wanted to know more about GUSEK and is happy to have Luiz answer some questions about GUSEK and his contribution to the field of Linear Programming and Operations Research.
What is your experience in Operations Research and/or Management Science and your background?
What is your experience in Operations Research and/or Management Science and your background?
I'm a Brazilian researcher and Master's student in the CPGEI department (Graduate School in Electrical Engineering and Industrial Computer Science) at UTFPR University. I'm actually working on two optimization projects to help in decision making. The first is focused in power distribution networks maintenance which is a partnership with COPEL (the energy company of Paraná State). The other one is a PETROBRAS (the world’s 7th biggest oil company) with oil distribution scheduling. The first project was my first contact with Operations Research, about a year ago.What is your experience with GLPK?
One COPEL project requirement is to use free software to solve LP/MILP models. After some research and a lot of tests GLPK was the natural choice: fast (not like commercial solvers, but better than other open source options), customizable, flexible, has its own modeling language (the GMPL) and is under constant development.Why did you want to create GUSEK?
Developing GMPL models without any resources (just with Windows Notepad) was a challenge. To make our job affordable and to save costs we made a custom IDE integrating the GLPK stand-alone solver and the GMPL programming language into a customizable text editor (SciTE). Custom improvements are made in the sense to be more friendly to the GMPL developer. After many changes, the IDE was significantly different from the original SciTE and we adopted the new name: Gusek (GLPK Under Scite Extended Kit).Who do you expect to use GUSEK?
People who want a free, easy-to-use interface for development and solving of LP/MILP problems, like GLPK users and students.Why did you want to create GUSEK for the Windows platform?
Commodity: in our lab we use Windows. But, as SciTE and GLPK are open source (and have Linux versions), GUSEK can be built for others OS, sure. Anyone which is available?Why did you distribute under the GPL license?
First, GLPK is a GNU project, released under GPL. SciTE has a similar license. Gusek was possible just because they are open source, so, why not open source? If anyone can use and contribute, Gusek will be a better tool for all. I've just started it and shared. GPL is a good choice to allow this.Are you working on any other free software projects?
Not really, actually I spend all of my time on the lab projects. However, me and my brother are planning to develop the open source concept of "cycle-tourism open tracks" based on our experience (take a look at www.odois.org, in Portuguese).Have you learned anything new or interesting while working on this project?
I'm involved with builds and patches to other people's software, learning with the open source community. I learned LUA and batch script languages to implement advanced functions into SciTE. I have learned GMPL and a lot of other LP/MILP modeling tools using Gusek, for sure!Thank you Luiz for this interview and sharing GUSEK with the Operations Research community. We hope that GUSEK finds a nice home in the Linear Programming world. We here at IEOR Tools have definitely added it to our tool bag.
Wednesday, February 4, 2009
Data mining is a necessary task for most Industrial Engineers and Operations Research analysts. Often times the analyst is forced to comb through a lot of data in order to gather information. There are many tools, often proprietary, to attack the piles of data. Some examples of great data mining tools include Excel or Access. Other more complicated tools include Microsfoft SQL Server or Oracle databases.
The Open Source community has been no stranger to developing tools for data mining. There are many free software equivalents for data mining for any analyst. Data can be mined in spreadsheets such as Openoffice Calc or KSpread. They can be compatible with windows files such as .xls for quick transfer from other data sources. These office suites are free to download on multiple operating systems.
Databases have been developed extensively in the open source community. Openoffice has an application similar to Access called Base. There are many free software database servers available for use including MySQL and PostgreSQL. There are numerous free software SQL clients including Knoda and GNOME-DB. PhpMyAdmin is a web-based MySQL client that can be setup with a webserver, such as Apache and interfaced with an internet browser.
There are also free software applications developed especially for data mining. These free software options include Weka and RapidMiner. These tools can be used as stand alone clients that can connect to any number of data souces, including proprietary ones.