Discussing the article: "Artificial Tribe Algorithm (ATA)"
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Check out the new article: Artificial Tribe Algorithm (ATA).
The ATA algorithm process starts with setting the parameters and randomly initializing the tribe, after which the fitness value is calculated. Next, the iteration counter is incremented and the current situation of the tribe is assessed. If the situation is favorable (the difference in the optimal fitness value between generations is greater than a given criterion), reproduction behavior is performed, where individuals exchange information. Otherwise, migratory behavior is used, in which individuals move based on the experience of both the individual and the entire tribe. Migration cannot be performed continuously to avoid excessive dispersion. The fitness value is then recalculated and compared with the best values recorded for the tribe and each individual. If a better solution is found, it is stored in memory. The termination conditions are checked and if they are met, the iteration is terminated. Otherwise, the process returns to the situation assessment step.
Including global information in ATA adds weight to the tribe's historical experience, helping to find better solutions and improve search capabilities. Increasing the weight of the tribe's experience helps improve the efficiency of the algorithm, accelerating convergence. To achieve this, ATA introduces a global inertial weight, which enhances search capabilities and speeds up the process.
The main innovation of ATA is the presence of a dual behavior system that adapts depending on the situation: reproduction is used for deep exploration when progress is good, and migration is activated when stuck in local optima, which promotes deeper exploration. The combination of individual and social learning is also important. Individual memory (Xs) is used during migration, and global memory (Xg) is weighted by the AT_w inertia factor. During the reproduction, partners are chosen randomly, which helps improve diversity and speed up the search.
Author: Andrey Dik