Hybrid Quantum-Behaved and Bio-Swarm Optimization Algorithm (BioQPSO)

Quantum-Behaved Particle Swarm Optimization (QPSO) of-fers strong global search ability but can suffer from reduced adapt-ability on certain multimodal and biologically structured search land-scapes. This paper proposes BioQPSO, a hybrid framework that inte-grates QPSO’s quantum-probabilistic position update model with bio-inspired swarm intelligence mechanisms drawn from ant colony and beeforaging behavior. Two concrete variants are introduced: AntBioQPSO,which incorporates pheromone-guided local leader selection and neigh-borhood communication, and BeeBioQPSO, which employs fitness-basedlocal leadership and scout-driven reinitialization to counter stagnation.Both variants extend the standard QPSO attractor by a third bio-inspiredcomponent with weight ϕ3, yielding a three-term weighted attractor thatbalances personal, global, and neighborhood guidance. The proposed al-gorithms are evaluated on ten standard benchmark functions (unimodaland multimodal, D = 30) over 30 independent runs with 30 particles and1000 iterations, and compared against PSO, QPSO, Ant Colony Opti-mization (ACO), and Artificial Bee Colony (ABC). Statistical validationusing Wilcoxon signed-rank and Friedman tests confirms significant per-formance differences across all benchmarks (p < 10−9 in all cases). In areal-world feature selection task on the Wisconsin Breast Cancer dataset(569 samples, 30 features), BeeBioQPSO achieves 94.83% classificationaccuracy using only 2 features on average, with the lowest standard de-viation (0.49%) among all tested algorithms. These results demonstratethat the BioQPSO framework provides a competitive and robust plat-form for complex optimization, with particular strengths on multimodalfunctions and practical feature selection tasks.