In comparison to its deterministic equivalents, simulation-driven meta-heuristic algorithms have proven effective in addressing a wide variety of issues. Despite this benefit, the stochastic nature of such algorithms results in a spectrum of answers after a certain number of trials, which may result in ambiguity about the quality of solutions. Therefore, it is critical to utilize an adequate method for evaluating the performance of a varied collection of meta-heuristic algorithms in order to make an informed judgment about the algorithms' supremacy and also to verify the claims made by researchers about their particular goals. Efficiency and effectiveness metrics may be used to evaluate the performance of a randomized meta-heuristic algorithm. The efficiency of an algorithm is related to the speed with which it finds correct solutions, converges, and computes. Effectiveness, on the other hand, refers to the algorithm's capacity for generating high-quality solutions. Both scopes are critical for solving continuous and discrete issues in single- or multi-objective settings. Each kind of issue has a unique formulation and measuring technique within the context of efficiency and effectiveness performance.
What is the best performing meta-heuristic optimization algorithm based on the mentioned measures?
Best depends on the problem and dataset. Meta-heuristics are always tuned for the problem and anyone can perform good for the said objectives of efficient and muture convergence. I can recommend GA because it is very flexible to fit for your problem.
There are many possible ways to utilize, hybridize and tune Meta-heuristics. I have disscussed some here https://www.researchgate.net/publication/353637329_Task_Scheduling_Optimization_in_Cloud_Computing?_sg=NCJB38yEx5UlUlkAJa1_EwyKr9zMQeeXrMP4B29178J4aPyjgV3f-z6YhwZvzCOL0LsYDf3hYIfVJmdM3UcqK5btt4n4UN-VC-WNdPqP.1QGCA9-7csqI5cmc2as0IoSH7xcs5B5KS5p1oG56itUK1_5C08G5rWJ6v29tfGZKHC7ipbnIaxB0EdnZuEfGMQ
One possible variant can be found in this article
https://www.researchgate.net/publication/353040925_Balancer_Genetic_Algorithm-A_Novel_Task_Scheduling_Optimization_Approach_in_Cloud_Computing?_sg=fyEwSWR9DrbDoy6hJmTMC2mt_kD7Bpopf4yOKOoTxMWAnbI-XK0qi4RU-R7Sk3KjpcT34e9Y2jQcaZjKa3V8_PnDd8i4XjwYvGJSUKjw.5DqKLELWKd1mR2h2ju6e2cLfMYhEsDRJ8mzeHfYhGu4M7ZQSGhF0x5pUL9DfO2GrIseI9TPcEPahh1Oehe5_WQ
Thank you
It depends on your problem and your fitness function. The problem is discrete or continuous?
You should see which meta-heuristic algorithm is more match to your problem.
We can not select one of the meta-heuristic algorithms as the best for all problems.
Maybe these articles help you:
https://www.sid.ir/en/journal/ViewPaper.aspx?id=494476
Article Meta-Heuristic Approaches for Solving Travelling Salesman Problem
In general, all heuristics and meta-heuristics are good in tuning parameters offline. So you have to choose which one can provide fast speed convergence to reach the optimal value and that depends on ur objective function. But these classic optimization techniques have also several drawbacks:-
For example, in linearized models: the objective function is designed according to the nominal transfer function of the plant considering no-load disturbance. Therefore, poor performance may occur at the moment of load disturbance and in the case of variations of the system parameters. Therefore, you need a tool that can help the optimization algorithms to be more sensitive to load perturbations and changing system parameters. you can use Balloon Effect for this purpose.
Here are some papers for more details:-
1. Dahab, Y. A., Abubakr, H., & Mohamed, T. H. (2020). Adaptive load frequency control of power systems using electro-search optimization supported by the balloon effect. IEEE Access, 8, 7408-7422.
2. Abubakr, H., Mohamed, T. H., Hussein, M. M., Guerrero, J. M., & Agundis-Tinajero, G. (2021). Adaptive frequency regulation strategy in multi-area microgrids including renewable energy and electric vehicles supported by virtual inertia. International Journal of Electrical Power & Energy Systems, 129, 106814.
Dear Prof. Bryar Hassan,
Dear honorable researchers,
- Since most real-world optimization problems have different needs and requirements that vary from industry to industry, there is no universal algorithm or approach that can be applied to every circumstance, and, therefore, it becomes a challenge to pick up the right algorithm that sufficiently suits these essentials.
- All metaheuristic optimization approaches are alike on average in terms of their performance. The extensive research studies in this field show that an algorithm may be the topmost choice for some norms of problems, but at the same, it may become to be the inferior selection for other types of problems.
The following two links may possibly helpful in this context:
- https://www.researchgate.net/post/Are_you_interested_in_Metaheuristic_Optimization_Algorithms
Article Sea Lion Optimization Algorithm for Solving the Maximum Flow Problem