Gray Wolf Optimizer — How It Can Be Used with Laptop Imaginative and prescient | by James Koh, PhD | Feb, 2024
That is the final a part of my sequence of nature-inspired articles. Earlier, I had talked about algorithms impressed by genetics, swarm, bees, and ants. As we speak, I’ll discuss wolves.
When a journal paper has a quotation depend spanning 5 figures, you realize there’s some critical enterprise happening. Gray Wolf Optimizer [1] (GWO) is one such instance.
Like Particle Swarm Optimization (PSO), Synthetic Bee Colony (ABC), and Ant Colony Optimization (ACO), GWO can be a meta-heuristic. Though there’s no mathematical ensures to the answer, it really works nicely in apply and doesn’t require any analytical information of the underlying downside. This enables us to question from a ‘blackbox’, and easily make use to the noticed outcomes to refine our resolution.
As talked about in my ACO article, all these in the end relate again to the elemental idea of explore-exploit trade-off. Why, then, are there so many various meta-heuristics?
Firstly, it’s as a result of researchers should publish papers. A superb a part of their job entails exploring issues from completely different angles and sharing the methods during which their findings result in advantages over current approaches. (Or as some would say, publishing papers to justify their salaries and search promotions. However let’s not get there.)
Secondly, it’s as a result of ‘No Free Lunch’ theorem [2] which the authors of GWO themselves talked about. Whereas that theorem was particularly saying there’s no free lunch for optimization algorithms, I feel it’s truthful to say that the identical is true for Knowledge Science on the whole. There isn’t a single final one-size-fits-all resolution, and we regularly should attempt completely different approaches to see what works.
Due to this fact, let’s proceed so as to add one more meta-heuristic to our toolbox. As a result of it by no means hurts to have one other device which could turn out to be useful in the future.
First, let’s think about a easy classification downside on photographs. A intelligent method is to make use of pre-trained deep neural networks as characteristic extractors, to transform…