Consider a practical problem that has a good solution using a systematic algorithm. A researcher tries to find a better solution, and created a new solution method. The two solutions are comparable, but the new method is complex, relatively unreliable, and memory space and processor energy consuming.
Do you consider the new method a contribution, or an appreciated effort, or a failure?
It depends upon the application and requirements of user. Suppose if someone is only concerned about the solution quality and not the hardware requirements then new method can be termed as appreciated effort. However, you need to look into the reliability aspect and make it better.
It depends. The "systematic algoriithm" may be slow when solving large instances, for example, and then it might be a good idea to take that new method and improve it. Or invent a third one that kills the other two. :-) As was mentioned above - it depends on the purpose. If the input is quite similar from one application to the next perhaps a streamlined code that works for just those instances is the best choice. However it will be difficult to know IF you have found the right tuning of the code - and that's a tough nut to crack.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Probability
- Probability Theory
- Reliability