Add Owl Search Algorithm (OSA)

README.md

@@ -3184,6 +3184,9 @@ All visualization examples: [Link](https://mealpy.readthedocs.io/en/latest/pages
 
 ### O
 
+* **OSA - Owl Search Algorithm**
+  * **OriginalOSA**: Jain, M., Maurya, S., Rani, A., & Singh, V. (2018). Owl search algorithm: A novel nature-inspired heuristic paradigm for global optimization. Journal of Intelligent & Fuzzy Systems, 34, 1573–1582.
+
 ### P
 
 * **PSO - Particle Swarm Optimization** 

mealpy/__init__.py

@@ -41,7 +41,7 @@
                           DMOA, DO, EHO, ESOA, FA, FFA, FFO, FOA, FOX, GJO, GOA, GTO, GWO, HBA, HGS, HHO, JA,
                           MFO, MGO, MPA, MRFO, MSA, NGO, NMRA, OOA, PFA, POA, PSO, SCSO, SeaHO, ServalOA, SFO,
                           SHO, SLO, SRSR, SSA, SSO, SSpiderA, SSpiderO, STO, TDO, TSO, WaOA, WOA, ZOA,
-                          EPC, SMO, SquirrelSA, FDO)
+                          EPC, SMO, SquirrelSA, FDO, OSA)
 from .system_based import AEO, GCO, WCA
 from .music_based import HS
 from .sota_based import LSHADEcnEpSin, IMODE

mealpy/swarm_based/OSA.py

@@ -0,0 +1,88 @@
+#!/usr/bin/env python
+# Created by "Furkan Buyukyozgat" at 15:25, 05/01/2026-------%
+#       Email: furkanbuyuky@gmail.com                        %
+#       Github: https://github.com/furkanbuyuky              %
+# -----------------------------------------------------------%
+
+import numpy as np
+from mealpy.optimizer import Optimizer
+
+
+class OriginalOSA(Optimizer):
+    """
+    Original Owl Search Algorithm (OSA).
+
+    Main steps:
+    - Normalize intensity using best/worst fitness
+    - Compute distance to best
+    - Intensity correction: ic = I / (r^2 + eps) + rand()
+    - Update position: x_new = x +/- beta * ic * |alpha * best - x|
+    - Apply bounds and greedy selection
+    """
+
+    def __init__(
+        self,
+        epoch=10000,
+        pop_size=100,
+        beta_max=1.9,
+        alpha_max=0.5,
+        eps=1e-12,
+        **kwargs
+    ):
+        super().__init__(**kwargs)
+        self.epoch = self.validator.check_int("epoch", epoch, [1, 100000])
+        self.pop_size = self.validator.check_int("pop_size", pop_size, [5, 100000])
+
+        self.beta_max = self.validator.check_float("beta_max", beta_max, (0.0, 10.0))
+        self.alpha_max = self.validator.check_float("alpha_max", alpha_max, (0.0, 1.0))
+        self.eps = float(eps)
+
+        self.set_parameters(["epoch", "pop_size", "beta_max", "alpha_max"])
+        self.sort_flag = False
+
+        self.alpha = None
+
+    def _beta(self, epoch: int) -> float:
+        iter_ = epoch + 1
+        return self.beta_max - self.beta_max * (iter_ / float(self.epoch))
+
+    def _get_worst_agent_current_pop(self):
+        fits = np.array([ag.target.fitness for ag in self.pop], dtype=float)
+        if self.problem.minmax == "min":
+            idx = int(np.argmax(fits))
+        else:
+            idx = int(np.argmin(fits))
+        return self.pop[idx]
+
+    def evolve(self, epoch: int) -> None:
+        if self.alpha is None:
+            self.alpha = self.generator.random() * self.alpha_max
+
+        pvm = self.generator.random()
+        beta = self._beta(epoch)
+
+        bestowl = self.g_best
+        weakowl = self._get_worst_agent_current_pop()
+
+        denom = (weakowl.target.fitness - bestowl.target.fitness)
+        if np.abs(denom) < self.eps:
+            denom = self.eps
+
+        for i in range(self.pop_size):
+            x = self.pop[i].solution
+
+            intensity = (self.pop[i].target.fitness - bestowl.target.fitness) / denom
+            r = np.linalg.norm(x - bestowl.solution)
+            ic = (intensity / (r * r + self.eps)) + self.generator.random()
+
+            step = beta * ic * np.abs(self.alpha * bestowl.solution - x)
+            if pvm < 0.5:
+                x_new = x + step
+            else:
+                x_new = x - step
+
+            x_new = self.correct_solution(x_new)
+            agent_new = self.generate_agent(x_new)
+
+            if self.compare_target(agent_new.target, self.pop[i].target, self.problem.minmax):
+                self.pop[i] = agent_new

tests/swarm_based/test_OSA.py

@@ -0,0 +1,33 @@
+#!/usr/bin/env python
+# Created by "Furkan Buyukyozgat" at 15:25, 05/01/2026-------%
+#       Email: furkanbuyuky@gmail.com                        %
+#       Github: https://github.com/furkanbuyuky              %
+# -----------------------------------------------------------%
+
+import numpy as np
+import pytest
+
+from mealpy import FloatVar, OSA, Optimizer
+
+
+@pytest.fixture(scope="module")
+def problem():
+    def objective_function(solution):
+        return np.sum(solution ** 2)
+
+    problem = {
+        "obj_func": objective_function,
+        "bounds": FloatVar(lb=[-10, -15, -4, -2, -8], ub=[10, 15, 12, 8, 20]),
+        "minmax": "min",
+        "log_to": None
+    }
+    return problem
+
+
+def test_OSA_results(problem):
+    model = OSA.OriginalOSA(epoch=10, pop_size=50, beta_max=1.9, alpha_max=0.5)
+    g_best = model.solve(problem)
+
+    assert isinstance(model, Optimizer)
+    assert isinstance(g_best.solution, np.ndarray)
+    assert len(g_best.solution) == len(model.problem.lb)