[FEAT] Add Dung Beetle Optimizer (DBO) to swarm_based#213
[FEAT] Add Dung Beetle Optimizer (DBO) to swarm_based#213ErenKayacilar wants to merge 4 commits intothieu1995:masterfrom
Conversation
This PR adds a new swarm-based optimizer: **Dung Beetle Optimizer (DBO)**. - Implemented `OriginalDBO` in `mealpy/swarm_based/DBO.py` - Exposed the algorithm in `mealpy/__init__.py` - The implementation follows the structure and coding style of existing swarm-based optimizers (e.g., COA, BFO). - Behavior is based on the original paper: Xue, J., & Shen, B. (2022). Dung beetle optimizer: A new meta-heuristic algorithm for global optimization. *The Journal of Supercomputing*, 79, 7305–7336. - Tested on a simple Sphere function (minimization) to verify convergence and correct integration with MEALPY.
thieu1995
left a comment
thieu1995
left a comment
There was a problem hiding this comment.
There are many more problems I can see in the code. I will update the code for better performance.
mealpy/swarm_based/DBO.py
Outdated
| Generate an empty agent (without target). | ||
|
|
||
| DBO does not require any additional attributes per agent, | ||
| therefore only the solution vector is stored. |
There was a problem hiding this comment.
You should delete this function. If you don't have any custom property in agent.
mealpy/swarm_based/DBO.py
Outdated
| epoch (int): The current iteration. | ||
| """ | ||
| # Make sure previous positions array is aligned with the population | ||
| if self._prev_positions is None or len(self._prev_positions) != self.pop_size: |
There was a problem hiding this comment.
No need for this condition. It will always the same size.
mealpy/swarm_based/DBO.py
Outdated
| [agent.solution.copy() for agent in self.pop] | ||
| ) | ||
|
|
||
| pop_array = np.array([agent.solution.copy() for agent in self.pop]) |
There was a problem hiding this comment.
No need for copy() operator here. Because you don't change the pop_array, you just use it.
mealpy/swarm_based/DBO.py
Outdated
| ) | ||
|
|
||
| pop_array = np.array([agent.solution.copy() for agent in self.pop]) | ||
| prev_array = self._prev_positions.copy() |
There was a problem hiding this comment.
Same for this array. In fact, you can just use 1 variable, defined it as self.prev_pos. No need to use copy() and extra prev_array (wasted memory).
| pop_array = np.array([agent.solution.copy() for agent in self.pop]) | ||
| prev_array = self._prev_positions.copy() | ||
|
|
||
| # Global best / worst positions (bestX and worstX in the paper) |
There was a problem hiding this comment.
Delete all of this block. You can use this function. https://github.com/thieu1995/mealpy/blob/master/mealpy/optimizer.py#L511
anh9895
left a comment
anh9895
left a comment
There was a problem hiding this comment.
Please fix all the comments.
mealpy/swarm_based/DBO.py
Outdated
| self.pop + pop_new, self.pop_size, self.problem.minmax | ||
| ) | ||
|
|
||
| # Update previous positions x(t−1) for the next iteration |
There was a problem hiding this comment.
This code should be appeared before the "Merge old and new populations".
If not, the self._prev_positions and self.pop is always the same.
mealpy/swarm_based/DBO.py
Outdated
| 79, 7305–7336. | ||
| """ | ||
|
|
||
| def __init__( |
There was a problem hiding this comment.
Write code inline, not break line like this please.
|
Hi @thieu1995 and @anh9895, First of all, I sincerely apologize for the long delay in my response. I had some personal/academic commitments that kept me away from this project for a while. Thank you so much for the detailed feedback and code reviews. I've carefully reviewed all your comments regarding memory optimization, code structure, and the positioning of the population updates. I am still very much interested in contributing the Dung Beetle Optimizer to Mealpy. I will start implementing the requested changes and resolve the merge conflicts as soon as possible. Thanks for your patience |
Removed custom generate_empty_agent() override since DBO has no additional agent properties; parent class method is sufficient. Removed unused Agent import. Removed unnecessary _prev_positions size check in evolve() — population size is always consistent. Removed redundant copy() calls on pop_array and g_best/g_worst since they are only read, not mutated. Eliminated extra prev_array variable; using self._prev_positions directly to avoid wasted memory. Replaced manual best/worst agent logic with built-in self.get_worst_agent() from the Optimizer base class. Moved _prev_positions update to before the merge step so it correctly captures the current population state, not the already-merged one. Reformatted init signature to inline style as per code style guidelines.
3 participants
This PR adds a new swarm-based optimizer: Dung Beetle Optimizer (DBO).
🔧 What’s Included
OriginalDBOinmealpy/swarm_based/DBO.pymealpy/__init__.pyCloses #
📑 Description
Adds an implementation of the Dung Beetle Optimizer inspired by the original paper and integrates it into the MEALPY swarm-based module. Ensures compatibility with the current optimizer interface, population handling structure, and evaluation pipeline.
✅ Checks
ℹ Additional Information
No breaking changes.
The implementation is self-contained and does not modify other algorithms.