Saurabh Samdani - Academics

Ant Colony Optimization

Many people ask me how I got so fond of this optimization technique called ant colony optimization. This is a brief anecdote about my interest in this topic. You might want to check out the Ant colony optimization Homepage

When I was in my second ( sophomore) year, I participated in a bridge design competition organized by Technfest IIT Bombay. The objective was to design a minimum cost bridge ( using a simple software) which can carry the given loads. Initially when I participated in the competition, we tried trial and error approach by varying the cross section areas, topology of the truss, the coordinates of the nodes and so on. We submitted our designs, but we could not win the competition.

So I started thinking what one should do to design a bridge better than the winning design. I thought of writing the cost function and differenciating it with respect to the variables to obtain the lowest cost. But in this problem there were more than 50 variables and none of them continuous. Since cross section areas are to be restricted to market available sizes the variables are discrete. So even if one wrote the objective function, how was I to differentiate it. I approached Dr. Arvind kumar Singh . to check whether he would help out. He flatly refused me saying that he is not interested in working with a Second year student on a research topic. I was disappointed but did not give up my quest to solve the problem. I met some seniors and particularly Ashwinikumar Patil helped me by giving me pointers to genetic algorithms and some other algorithms.

Then my plan was to work on this topic in summer holidays. I started preparing hard for my end semester examinations and kept this work at abeyance for some days. It was my habit to study in the Central Library, it offered a great environment for studying. Then one day I serendipitously came across a book titled Ant Algorithms which was a conference proceeding with picture of ants on the cover page. I picked it up out of mere curiousity, as to know what abstruse topic this would be called "Ant Algorithms". I perused through some of the papers, and then it striked me that I could use this technique to perform discrete optimization in design of the truss. After end semester exams, I worked hard on this topic with Vinay Belambe and Ashwninkumar Patil and successfully solved the problem.

We wrote a paper on this topic and it was presented at the International Conference on World of Innovations in Structural Engineering, Hyderabad Dec.1-3,2004. Our work won the BEST PAPER AWARD.

''Discrete optimization of Trusses using Ant colony metaphor.''
Saurabh Arun Samdani,Vinay Belambe, Ashwinikumar Patil.
( Best Paper Award at International Conference on World of Innovations in Structural Engineering, Hyderabad Dec.1-3,2004. ). Download: here [pdf format, 412KB]

Now I have generalized the code I used to optimize the trusses. Please check out the Source Codes This code can be used for any discrete optimization problem, most typically the NLP problems.

My Publications

''Site Layout of Temparory Construction Facilities using Ant Colony Optimization.''
Saurabh Arun Samdani, Lalit Bhakal, Arbind Kumar Singh.
( Accepted for presentation at ASCE 4th International Engineering and Construction Conference, July 28 2006, Los Angeles, CA ).
Acceptance Letter: click here.
''Ant Colony Optimization in Earthwork Allocation.''
Suresh Kataria, Saurabh Arun Samdani and Arvind Kumar Singh
( Accepted for presentation at International Conference on Intelligent Systems, December 1-3 2005 Kuala Lumpur, Malaysia ).

Acceptance Letter: click here.
''Discrete optimization of Trusses using Ant colony metaphor.''
Saurabh Arun Samdani,Vinay Belambe, Ashwinikumar Patil.
( Best Paper Award at International Conference on World of Innovations in Structural Engineering, Hyderabad Dec.1-3,2004. ). Download: here [pdf format, 412KB]


	\| home \| Source Codes \| academics \| resume \| friends \| family \| photos
	Ant Colony Optimization Many people ask me how I got so fond of this optimization technique called ant colony optimization. This is a brief anecdote about my interest in this topic. You might want to check out the Ant colony optimization Homepage When I was in my second ( sophomore) year, I participated in a bridge design competition organized by Technfest IIT Bombay. The objective was to design a minimum cost bridge ( using a simple software) which can carry the given loads. Initially when I participated in the competition, we tried trial and error approach by varying the cross section areas, topology of the truss, the coordinates of the nodes and so on. We submitted our designs, but we could not win the competition. So I started thinking what one should do to design a bridge better than the winning design. I thought of writing the cost function and differenciating it with respect to the variables to obtain the lowest cost. But in this problem there were more than 50 variables and none of them continuous. Since cross section areas are to be restricted to market available sizes the variables are discrete. So even if one wrote the objective function, how was I to differentiate it. I approached Dr. Arvind kumar Singh . to check whether he would help out. He flatly refused me saying that he is not interested in working with a Second year student on a research topic. I was disappointed but did not give up my quest to solve the problem. I met some seniors and particularly Ashwinikumar Patil helped me by giving me pointers to genetic algorithms and some other algorithms. Then my plan was to work on this topic in summer holidays. I started preparing hard for my end semester examinations and kept this work at abeyance for some days. It was my habit to study in the Central Library, it offered a great environment for studying. Then one day I serendipitously came across a book titled Ant Algorithms which was a conference proceeding with picture of ants on the cover page. I picked it up out of mere curiousity, as to know what abstruse topic this would be called "Ant Algorithms". I perused through some of the papers, and then it striked me that I could use this technique to perform discrete optimization in design of the truss. After end semester exams, I worked hard on this topic with Vinay Belambe and Ashwninkumar Patil and successfully solved the problem. We wrote a paper on this topic and it was presented at the International Conference on World of Innovations in Structural Engineering, Hyderabad Dec.1-3,2004. Our work won the BEST PAPER AWARD. ''Discrete optimization of Trusses using Ant colony metaphor.'' Saurabh Arun Samdani,Vinay Belambe, Ashwinikumar Patil. ( Best Paper Award at International Conference on World of Innovations in Structural Engineering, Hyderabad Dec.1-3,2004. ). Download: here [pdf format, 412KB] Now I have generalized the code I used to optimize the trusses. Please check out the Source Codes This code can be used for any discrete optimization problem, most typically the NLP problems. My Publications ''Site Layout of Temparory Construction Facilities using Ant Colony Optimization.'' Saurabh Arun Samdani, Lalit Bhakal, Arbind Kumar Singh. ( Accepted for presentation at ASCE 4th International Engineering and Construction Conference, July 28 2006, Los Angeles, CA ). Acceptance Letter: click here. ''Ant Colony Optimization in Earthwork Allocation.'' Suresh Kataria, Saurabh Arun Samdani and Arvind Kumar Singh ( Accepted for presentation at International Conference on Intelligent Systems, December 1-3 2005 Kuala Lumpur, Malaysia ). Acceptance Letter: click here. ''Discrete optimization of Trusses using Ant colony metaphor.'' Saurabh Arun Samdani,Vinay Belambe, Ashwinikumar Patil. ( Best Paper Award at International Conference on World of Innovations in Structural Engineering, Hyderabad Dec.1-3,2004. ). Download: here [pdf format, 412KB]