In my scientific research, I use mainly the fuzzy sets theory. Now, I use the fuzzy logic into Multi-Criteria Decision-Making problems. What is your experience with artificial intelligence methods?
ANN and GA but in general that depends on the problem. Define your problem and check whether a specific AI method can solve it . In addition, if many AI methods can solve it check which is more critical to you accuracy or speed. Some time both can be achieved and that depend on the problem again.
I dont use AI much these days. But earlier i used to work on Neural Network. I just got hold of them it was easier for me to work through Neural Networks.. I know they are not the best algorithm..
I prefer Case Based Reasoning (CBR) to ANN. It requires less training and it is actually easier to check on what's happening inside. However, you need to be very careful designing the problem instance and the adaptation algorithm.
I am doing researches on evolutionary algorithms (EAs) and its applications on resource optimization such as production scheduling and vehicle routing. EAs (and other metaheuristics) are flexible algorithm frameworks and can be applied to solve many kinds of optimization problems.
For application-wise researches, we focus on designing problem-specific algorithm components including crossover and mutation. We try to understand the target problem well and incorporate domain knowledge into the algorithm to improve the solution quality.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence