Network Plot with Gephi - My 1st Trial

After quite sometime from my initial blog about Gephi, finally I took some time to learn using it.  Actually, for few days I've been trying to find some available network plot packages that can be developed within R environment.  However I can't find one that can produce appealing visualization better than Gephi (at least for now).  For those of you who'd love to learn making network plots using R, I do recommend Dai Shizuka's blog for clear starter tutorials.  For those who want to learn Gephi, it will take you about one or two days to make descent graphs.  Gephi is quite easy to learn for a beginner.  In principle, I only need to read two guides (i.e. the quick start guide and the supported input files) to start.  For making a try out plot (below) I spent most of the time in preparing the edge list input file.  So give some time to try it yourself then... :)     

My First Gephi Plot.

5 Abstraction Levels for Analyzing Organizations

Citing Bakos (1987, p22), "Organizations can be viewed at different levels of analysis ... :

• an individual performing a task.
• a work group including many individuals.
• an organization consisting of several groups.
• an industry with a number of firms.
• the entire economy, or society as a whole. " 

community by russel, openclipart.org

Reference:

Bakos, J. Y. (1987).  Interorganizational Information Systems: Strategic Opportunities for Competition and Cooperation.  PhD Thesis.  Massachusetts Institute of Technology.

2 Guru Videos on Multi-agent Coordination

For those who are interested in multi-agent coordination, its very possible that you may have "touched" Ed Durfee's paper in IEEE Computer (i.e. "Scaling Up Agent Coordination Strategies") and the very nice overview paper by Ed Durfee and Victor Lesser that was published in IEEE TKDE (i.e. "Trends in Cooperative Distributed Problem Solving").


I have tried to find some inspirations and found 2 videos on multi-agent coordination that were presented by the Ed Durfee and Vic Lesser.  Gr8T!!!


The following video is presented by Ed Durfee. 



For more details, two extra videos (presented by his PhDs) that explain the same topic can be seen via these videolectures.net url: the "Influence-based Policy Abstraction for Weakly-coupled DEC-POMDPs" and "Algorithms for Solving the Multiagent Simple Temporal Problem".   


For Victor Lesser's video, follow this video lecture. net's link.

videolectures.net

Have fun.. May the inspirations be with you.. :D  ho.. ho.. ho..

Barabási's "Linked" + "Bursts" in 1 Hour Video

videolectures.net

 

Follow this videolectures.net link for the video.  Unfortunately, it seems that the presentation for the "bursts" part is truncated.. Too bad, I could not get the full access to the section that I needed the most.. =(

"The Evolution of Cooperation" Review

Hoahh.. Finally, I've managed to finish Axelrod's classical book, The Evolution of Cooperation. I've borrowed the book from the company's library several months ago. The book is about how cooperation can emerge in the world of self interested agents. Based on 2 series of iterated prisoner dilemma competitions, it is shown that the unsophisticated TIT for TAT strategy is able to defeat its competing strategies, which often utilize more complicated approaches. The tactical characterization of the TIT for TAT simple strategy is then explained, generalized to a broader context (such as the world war 1 and the field of evolutionary biology), and compiled into cooperation strategies advises. For me, the first 3 chapters (out of 9 chapters) are the most interesting ones. The rest is rather lengthy elaboration for the first three chapters' pointers. The lesson offered is definitely an eye opener, the "story telling" style is not for me though. Anyway,the book is a must have for my private library and most importantly I've read it completely yeaa.. :p

Nice Introduction to Metaheuristics (E-Book)

Hi..
Trough all these times, I have refused to invest some time to understand some concepts that a lot of people may consider as trivial. When many people/papers mention some metaheuristics buzz words like: "local optima", "hill climbing", "tabu search", "simulated annealing", "swarm optimization", "genetic algorithms", etc; it just freaked me out. I don't think I have enough time to drill down those academic literatures, which is too formal for a brief introduction.

However, I have found an e-book from Sean Luke of GMU. The "Essentials of Metaheuristics" book, that he wrote, is a perfect starter for beginners. It is presented in a rather casual (informal) language, which makes the reading less boring. Each concept is always presented with the accompanying pseudocode (great!!!). I am really satisfied with the eBook. I am considering buying the hardcopy also.

www.openclipart.org

Surprisingly, Sean Luke is also the designer of an agent based simulation platform called MASON. Interesting.. 

I am still using Netlogo for making simulation and I find that the community (mailinglist) is really active and reliable. Based on my personal experience, there is always someone in the forum who will reply to your questions, even to the stupid ones (may be not too stupid :p). This is waaaay different with the Repast's (agent based simulation platform) community which is (basically) dead.

So the lesson is..: you should consider of building a live community to support the use of your platform. Great product with dead community has a great "likelihood" to die soon..

Once again, it is a great book, I finally understand that all those buzz words i mentioned before, have the same principle (the survival of the fittest). O yeah.. Sean Luke also mentioned another alternative e-book that you can use in learning the optimization heuristic topics: Global Optimization Algorithms: Theory and Application by Thomas Wiese. I never read it, but I may do it if the time permits.

OK have fun then..

Reduced Image Quality when Converting "Word" File to Pdf File

Hi All...

Some of you may have trapped in the same problem like I did. When I tried to convert a MS Word file to a Pdf file with some free converter, some of the images in the file were downgraded in quality (see the screen shot below).

If you have the same problem, you can use the free plug in from the microsoft dowload center. The plug in name is "2007 Microsoft Office Add-in: Microsoft Save as PDF". You can see the result in the picture shown below.

OK, have fun then... ;)

Desirability Index - Multiobjective Optimization

Just want to sketch my personal (graphical) interpretation of the desirability index approach for multi-objectives optimization problem.

Reference:
Derringer, G., & Suich, R. (1980). SIMULTANEOUS-OPTIMIZATION OF SEVERAL RESPONSE VARIABLES. Journal of Quality Technology, 12(4), 214-219.

Netlogo runs faster in windows than in ubuntu - Lesson learned from simulation try outs

Hi Everyone..


It's been quite a while since the last blog posting... This weekend I have been busy doing Netlogo simulation's try outs. Since I have 2 machines (1 old desktop and 1 laptop) in my home, I'd like to know which one will bring better performance in terms of computation speed. Moreover, I have an access to a remote machine (server) from a good colleague (lets call him "Master of Ninja Turtles", hahaha..). However, based on the try outs run in my local machines, I decided not to continue the try out using the remote machine. This performance test will be useful in determining the machine I should use in doing future simulation runs.. The following picture portrays the result, sorry about the censors.. hehehe..

The result of the performance test is shown in the figure below. 3 machine setups are tested for running the exact simulation scenario. As shown, based on my experience, the Netlogo simulation duration explodes when we use ubuntu. It seems that Netlogo is optimized for windows OS and not for ubuntu OS.

Here are some chat records that I extracted from the discussion (about the result) that I had with the "Master of Ninja Turtles".

Me: "before i thought that everything will run faster on linux... but its not always the case or?"

Master: "no, it really depends on the software, if I write something and optimise it on Linux, it might run slower on Windows. And vice versa"

Me: "cool master... i learned something... "

Master: "Windows has the advantage that there are more users, so companies usually create the application for windows, and support Linux as well. But they focus on windows. (makes sense to do something that benefits most of your customers) and in this case I guess the sun vm would also make a difference, but I don't know how much"


OK Have a productive week everyone...!!!

Netlogo Introductory Tutorial Video - Gabriel Wuerzer

What a coincidence, for my 110th blog, I post this beautiful Netlogo (agent-based simulation platform) tutorial created by Gabriel Wuerzer of Vienna Technical University. I found this (1 Hour 48 Minutes) video  from Youtube, and I think it is good to repost the news. There is not so many researcher who is willing to invest some time for sharing their knowledge. That is why this video is so valuable, especially for anyone who is interested in studying agent-based simulation.



I have watched the first 37 minutes of the video. I find the video easy to follow and really "beginner friendly". Not just presenting one way explanation, the video will guide you for doing simple "hands-on" tutorials also. The first 37 minutes focus on the concept of turtle (the agent). The pace of the lecture is quite slow, which is good for beginners. Can't wait to continue... :)

Concept of a Turtle (Agent) by Gabriel Wuerzer

01012011 - Linked Review

Finally.. I manage to finish the "Linked-Barabasi" book. Took me quite a while to finish the book, since some urgent tasks "busted" into my priority list LOL.


"Linked" was published earlier than "Bursts", however I read them in an adversed way. Personally, I find that Linked is more enjoyable to read than Bursts. Although Linked was firstly published 7 years ago (2003), it is contextual to our daily life (the era of internet). In Bursts, I pretty much annoyed by the papacy chapters. In Linked, some accompanying stories are perfectly coupled with the concepts introductions.

I learned many new ideas from Linked. The dispute of Erdös-Renyi network, the scale free network topology, the six degrees of separations concepts are some of them. While I was reading, I list some of the concepts that have caught my attention. I attach the sketch below. I think that the way Barabasi explains the concept of network thinking is really good.


Some years ago a guy said, "It is good to read something that is written by an expert from different school of thought". I think that his opinion makes sense. A person living in a box is able to see the interior but not the exterior. Hence having another perspective will build a more holistic view. Anyway, have a nice weekend, and have a brilliant new year too :p

Barabasi Bursts's Video @ Google Talk

Great..!!!
I found another Barabasi's video from Google about Bursts concept. I think that this video is better that the one I posted before, since it has more discussions and it encapsulates the content of the book better. High predictability of social occurences concept that Barabasi has forwarded, relies mostly on big chunk of historical data utilization which now can be operationalized by applying data mining/ machine learning methods.



One main character discussed in the Bursts book is Hasan Elahi. When I tried to search for his videos, I can easily find one. I post one of them which I find interesting. He managed to give the impression that he is not a boring guy :p. It always good to connect a bit more to the characters in the book that I have read. It brings more understanding and more connection to the idea. Nice weekend all..

Cool Lecture Videos: Introduction to Multiagent Systems

I think its really cool that a Professor would spend some time giving lectures which are accessible for everyone by using the web. While studying the multiagent system (MAS) subject, I found a set of lectures uploaded on youtube by one of the Gurus in the MAS, Michael Wooldridge of University of Liverpool. Its just great to feel his enthusiasm in delivering the talk.



If you browse down his youtube channel, you will be able to browse all videos. Starting from video 02-06 the formalization of the MAS discussion starts. The videos with lower serial numbers are more conceptual. It just brilliant. For me learning from videos is more effective, more time efficient, and more fun. You can fast forward the things you already understand, and you can rewind the things that you find difficult to grasp. It is just awesome. I think more lectures should be uploaded more to the internet. Global students should be able to learn from global professors and be mesmerized :p Anyway, I embed also an interesting video uploaded by Prof. Wooldridge. Enjoy :)

Video - Barabasi Talk: Human Mobility Patterns

Just found his (Barabasi's) talk from one of my favourite sites, the videolectures.net. Quite cool to see him giving a talk about the bursts phenomenon. Folow this link and Enjoy ;)

Hans's Thesis Review

Since in the near future, I will conduct a Multi Agent Systems (MAS) based research in the container terminal domain, I read Hans's thesis as part of my literatures review. He developed MAS based Decision Support Systems (DSSs), each for transportation load allocation case and barge rotation case. Personally, Hans put the strong emphasis on the "real time" benefit of using MAS-DSS instead of centralized DSS. The flexibility and the robustness of MAS DSS in coping with real time disruptions are the main points.


The thesis reflects intense communications between the researchers and the practitioners. I think the narration could have been worked a bit more so that the thesis could be digested better. Well just a personal thought though, its a great work anyway..

Coordination: Challenges & Enhancement Proposals

Challenges:
1. Cost and benefits unbalanced load
2. Low motivation
3. Strategic considerations of the members
4. Lack of dominant firm
5. Risk aversion attitude
6. Short-term benefit rather than long-term benefit orientation


Enhancement Proposals:
1. Incentives introduction
2. Inter-firm alliances creation
3. Scope resizing
4. Collective action creation.

(Van der Horst & De Langen 2008; Moonen 2009)

Lessons Learned from Bursts...

Randomness is disputed ...

Prioritization is the key ...

Bursts forward the generalizability of occurences ...

Personally, I don't like the lengthy crusades stories though ...

Information Management Top Journals List

Hi, I just want to forward my journal inventory search for those of you who also do research in the Information Management branch. Actually I just repost the journals list which is considered the top list in our school, the ERIM Journals List. Since in the web site the information management top journals list is combined with the logistics management journals list, I guess my simple work here might be useful for you. In our school, the STAR journals are considered the best ones, The P journals are the second best, and the S journals are the third best. Hope it will be useful.

The STAR Journals:

The P Journals:

The S Journals:

Auction Theory and Auction Practice Discrepancy - a Paper Review

Today I presented a paper review in LARGE group. The paper is from Management Science (1994) entitled “Modeling Competitive Bidding: a Critical Essay” written by Michel Rothkopf (RIP) and Ron Harstad. These 2 big economists in auction area write an essay which explains that there is big discrepancy between the existing (though its 1994 paper, I think it’s still relevant to the current research context) auction theory and the real practice, so that the developed theory is having limited application to the real live condition. My main interpretation of the whole paper is bundled by the figure shown below:


From this paper, we learn that the theories which try to model the competitive bidding behavior in auctions (both decision theory and game theoretical model) are mostly written in a single and isolated auction context. Moreover the assumptions which frame the theories (e.g. single isolated auctions, fixed number of bidders, symmetric Nash equilibrium, risk neutral bidders, etc) also limit the real world applicability. They suggest the researchers to adjust and enrich the elegant but simplified existing models so that the bidding model could be utilized in the real world situations.

There were a was discussion on how should we develop an analytic model which will fit to our research interest, which is the Dutch Flower Auction. Our research context is unique, since we are dealing with multi unit, multi bidders, multi suppliers, multi attributes, and highly interdependent auctions. We agreed that this paper would be considered as our main reference since it brings a lot of insights on how to make a realistic an applicable model, not just an elegant model which is good for theoretical sake but could not be applied in reality.

This writing also available in LARGE blog.