Computing-Inspired Algorithms in Biological Big Data Cloud Computing Platforms

Fengrong Xu
School of Art, Hanjiang Normol University, Shi Yan, 442000, China

Abstract:

In response to the computational bottlenecks and parallelization problems caused by high-dimensional heterogeneity in biological big data, this study introduces QPSO (Quantum-inspired Particle Swarm Optimization) and QA (Quantum Annealing). QPSO enhances global search capabilities through quantum superposition to solve local optimal problems, while QA accelerates optimization convergence, improves the parallel processing efficiency of the cloud platform, and optimizes data analysis performance. In the initialization phase, the quantum state encodes the particle position, and the quantum potential function is calculated to update the particle group position. During the update process, the quantum superposition calculation global attractor is introduced to adjust the particle search range, and the optimal solution is selected based on the measurement probability. The Hamiltonian of the target optimization problem is constructed, and the problem is converted into a quantum state evolution process. By adjusting the system temperature parameters and using the tunnel effect to find the global optimal solution, the energy function is dynamically adjusted during the cooling process to make the solution converge to the optimal state. On the cloud computing platform, the particle swarm of QPSO is divided into multiple sub-groups and assigned to different computing nodes to parallelly calculate the fitness value. The QA algorithm adopts a distributed simulated quantum annealing mechanism. Each computing unit independently executes the Hamiltonian update and synchronizes the optimization state through global communication. Based on the task scheduling strategy, computing resources are dynamically allocated to improve the utilization of computing nodes. Combined with the load balancing mechanism, task allocation is adjusted dynamically to ensure reasonable scheduling of computing resources. Experimental results show that in 30 independent experiments, the median MAE (Mean Absolute Error) of QPSO+QA is 0.12, and the median MSE (Mean Square Error) is 0.08, which is more accurate in biological big data optimization tasks. In the processing of biomedical big data, the execution time of the QPSO+QA algorithm increased from 0.5 seconds (  feature) to 48.7 seconds (  feature), an increase of 97.4 times, significantly lower than other algorithms.