 |
FROM IlliGAL (Illinois Genetic Algorithms Laboratory at University of Illinois) to KanGAL (Kanpur Genetic Algorithms Laboratory at IIT Kanpur), it has been a long journey for IIT Professor Kalyanmoy Deb.
Q. Your prize bio says you accidentally stumbled upon Evolutionary Algorithm. How?
A. Back in 1985, when I graduated from IIT-Kharagpur, most graduates wanted to settle down with plum job offers. I too joined a design company in Delhi. However, by the second year I started realising that this is not what I would like to do my entire life. I went to University of Alabama in USA.
At the university while glancing through various electives, I saw this course being offered by Professor David Bolger on Genetic Algorithms in Engineering Mechanics Department. They claimed to mimic natural evolutionary process to solve engineering problems. I had always been fascinated by evolution as a concept and picked up the course.
Q. How has the field evolved hence?
A. Although the idea of Genetic Algorithms was there since 1965, progress was immensely slow in the field and there was lack of much knowledge about it. But by the end of the 80s and early 90s, many people saw it as an alternative means to solve problems. It is difficult to use classical mathematics to solve non-linear problems of engineering. While pursuing my PhD, I came across fundamental issues in the field that needed to be looked upon. In 1992 I finished my PhD and returned to IIT Kanpur as an Assistant Professor.
Q. Were you always interested in finishing your PhD and returning?
A. Yes I was always interested in returning. I was working on a project after finishing my PhD for 9 months and then joined IIT-Kanpur. IIT-Kanpur gave me the freedom to pursue research in Genetic Algorithms, which was still at a nascent stage. IIT-Kanpur encouraged me and I continued here with some very good students.
|
|
|
|
| |
“To put it simply, many times we look for solutions that work and if certain designs satisfy those conditions, then it’s fine. But here we are looking at optimization, which means that not only are we looking at something that works but rather something that can’t be better” |
|
|
|
|
|
|
Q. Are competing efficiency concerns that require an optimal solution very relevant in the social sector?
A. Yes they are. They are needed in any branch of science, technology or commerce. To put it simply, many times we look for solutions that work and if certain designs satisfy those conditions, then it’s fine. But here we are looking at optimization, which means that not only are we looking at something that works but rather something that can’t be better.
So optimization can be used for any decision-making issue, management, engineering or manufacturing process that need to be done in the shortest possible time or in cheapest means possible. It has a purpose. These computerised algorithms provide solutions that can’t be bettered by anybody else. So these are useful tools.
Q. Has India moved up on the optimization efficiency curve?
A. Not really. There are not many Indian industries that are doing original research work in the arena. We need to produce indigenous products that are going to lead the world market. In India, we need to think on such lines and bring a change in ideas. As of now, most of my optimization ideas are being implemented by software firms and have their products based on them. They are selling these products to hard core engineering sectors like chemical industry, automobiles and aerospace. So indirectly, they are using my ideas through the software industries.
Q. What does the word ‘ genetic’ really signify in the phrase ‘Genetic Algorithm’?
A. There has been a lot of controversy regarding how ‘genetic’ are we when we use the word genetic algorithm. So I minimised usage of the word and instead use Evolutionary algorithm. Actually that is the correct term. Millions of years ago there were single-cell organisms. From there through evolution, we now have complex organisms that can adapt to various situations. This entire evolutionary process has been based on selecting the good traits and weeding out the bad ones.
This evolutionary process can also be used to manufacture a product/create a design that is more adaptive and generates profit. So there is similarity in the process. We are trying to use computer algorithms to create more adaptive artificial things. John Holland used the word ‘genetic algorithm’ first in 1970s. He was carrying out bit more genetic operations. But we are not exactly using his methods. So I prefer using Evolutionary Algorithm.
|
|
|
|
| |
“Jobs remain the most important criteria for gauging a student’s success. Getting a coveted job offer is considered over and above pursuing research or PhD. Thus, it is difficult to transgress to a research-oriented institution unless mindset changes” |
|
|
|
|
|
|
Q. IITs were conceptualized as engineering teaching institutions, and the aim was to produce good engineers and not exactly scientists/researchers. How much have we moved from that paradigm?
A. Personally, I feel it has been a very slow movement. I have been a part of the IIT system for 24 years now (four as a student and 20 as a teacher). The primary reason for this is the societal mindset. Jobs remain the most important criteria for gauging a student’s success.
Getting a coveted job offer is considered over and above pursuing research or PhD. Thus, it is difficult to transgress to a research-oriented institution unless the mindset changes. Outside India, there are strong industry-university collaborations. Many of the research ideas are funded by industries, nurtured in universities and take-off from there. I don’t see that happening in a major manner in India.
Q. Nowadays almost all researchers focus on patenting. What do you have to say about the sudden craze for patents?
A. I am totally against patents. I have never patented any of my ideas. If you patent something, you are prohibiting people to use it in a free manner. One has to either pay a large amount or follow certain conditions and then only they can use the ideas. I doubt that any of my ideas would have been used that much, had they been patented. In fact, my supervisor told me, you would have been a millionaire, if you had patented NSGA. But then it would not have been so much popular. Most of the universities are asking students to do more research, handle more projects and gain patents. It gives more fame and popularity to the institute.
You can’t blame the institutes for encouraging more patents. They are competing amongst each other to get a better ranking. But my idea of contribution to research in society is not based on patents. If you have an interesting idea and it is not being used by people because it is patented, then your idea is not successful. So I feel that universities should not focus on patenting.
