Ecological Cycle Optimizer: Population Search, Decomposition, and Greedy Selection
Summary
The article explains the Ecological Cycle Optimizer (ECO), a population-based metaheuristic that models candidate solutions as ecosystem roles. The population is divided into producers, herbivores, carnivores, and omnivores: the best candidates act as information sources, while the other groups move using selected members of the population as targets. A time-varying predation coefficient encourages broader movement early in the search and more cautious movement later.
After these group movements, ECO applies one of three decomposition procedures: movement near the best solution, a random local step scaled by distance to that solution, or a global random move. Greedy revision accepts a candidate only when its objective value improves, and out-of-bounds candidates are reset randomly within the search limits. The article discusses tests against other optimization algorithms on benchmark functions and reports strengths on some problem types, with weaker performance on discrete functions and substantial sorting and copying costs. These are general optimization tests, not evidence of trading performance; the article notes that some compared algorithms were modified and results depend on its experiments.
Key ideas
- ECO divides candidate solutions into four groups that share information through different movement rules.
- A decreasing predation coefficient shifts the search from larger exploratory moves toward local refinement.
- Three decomposition modes combine movement toward the best solution with local and global random exploration.
- Greedy revision preserves or improves the best objective value, while boundary violations are reset randomly.
- Benchmark results suggest problem-dependent performance, with difficulty on discrete functions and added computational cost.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.