According to a colleague working on optimization algorithms and different papers, Grasshopper optimization algorithm (GOA) performs better than PSO. In my case for a multiobjective problem, PSO performs better around 90% of runs than GOA. My verdict is PSO can perform better too depending on the nature of the problem. I need suggestion from experts/ seniors please....
yes, it can perform way better than the new metaphor based meta-heuristics. PSO and GA are well accepted among the research community.
The answer is yes. According to No Free Lunch Theorem (Wolper and Macready, 1997), the mean performance of two search algorithms (let's call them A and B) will be identical when using them to solve all possible optimization problems. This means that in some problems algorithm A will perform better than B and in others vice versa. That result is well known, or at least it should be. For that reason I recommend you not to take seriously people who tell you that their algorithm is a "superior" proposition to all other existing algorithms.
Wolpert, D. H. and Macready, W. G. (1997) ‘No free lunch theorems for optimization’, IEEE Transactions on Evolutionary Computation, 1(1), pp. 67–82. doi: 10.1109/4235.585893.
Of course it depends on the problem at hand.
Suggestion: try on a good benchmark with quasi-real world problems (for example CEC2011) and plot the CDF (Cumulative distribution functions of the final errors, over say 100 runs). If the curve for GOA is completely "above" the one for PSO, then, in that precise case GOA is better, and vice versa. If the curves cross no superiority can be claimed.
Note that according to the original position update formula 7 (https://www.baeldung.com/cs/grasshopper-optimization-algorithm) the mathematical consistency of GOA is questionable: the formula combines displacement with forces.
The formula 8 is more consistent, but is then just a variant of several already published methods.
BTW you may read
Sörensen, Kenneth. ‘Metaheuristics—the Metaphor Exposed’. International Transactions in Operational Research 22, no. 1 (2015): 3–18. https://doi.org/10.1111/itor.12001.
Camacho Villalón, Christian Leonardo, Thomas Stützle, and Marco Dorigo. ‘Grey Wolf, Firefly and Bat Algorithms: Three Widespread Algorithms That Do Not Contain Any Novelty’. In Swarm Intelligence, edited by Marco Dorigo, Thomas Stützle, Maria J. Blesa, Christian Blum, Heiko Hamann, Mary Katherine Heinrich, and Volker Strobel, 121–33. Lecture Notes in Computer Science. Cham: Springer International Publishing, 2020. https://doi.org/10.1007/978-3-030-60376-2_10.
Campelo, Felipe, and Claus Aranha. ‘Sharks, Zombies and Volleyball: Lessons from the Evolutionary Computation Bestiary: LIFELIKE Computing Systems Workshop 2021’. Proceedings of the LIFELIKE Computing Systems Workshop 2021, CEUR Workshop Proceedings, 3007 (13 November 2021). https://lifelikecs.organic-computing.de/.
And a nice (incomplete!) bestiary
http://fcampelo.github.io/EC-Bestiary/
Thank you Rohail Gulbaz, Maurice Clerc and Luis Velasco for the guidance. Glad that I am taking them correctly...Btw liked the phrase of Luis Velasco "I recommend you not to take seriously people who tell you that their algorithm is a "superior", haha....
- Badges
- Science topic
- Similar topics
- Mathematics