Finding optimal meta-heuristic parameters is one of the open problems in computation science today. However, works in the literature using Design of Experiments or the so-called Hyper-heuristics can be seen, which are meta-heuristics specialized in optimizing meta-heuristics. What methods do you know? Which do you think is best?
You can check these.
Article Parameter tuning for meta-heuristics
Article A Hybrid Metaheuristics Parameter Tuning Approach for Schedu...
Thanks.
Usually parameters are tuned experimentally based on the experience and problem at hand.
Optimizing parameters of Meta-heuristic using another Meta-heuristic is not ideally opted because of the computational cost.
Parameters can be made variable or adaptive. For example in GA mutation rate can be increased if deviation of fitness value is reducing and vice versa.
We have recently published two papers and proposed different adaptive methods for the hyperparameters tuning as follows
1. Article A deep learning-based evolutionary model for short-term wind...
Article Wind Turbine Power output Prediction Using a New Hybrid Neur...
- Badges
- Science topic
- Similar topics
- Books