High-performance language parallel database processing learning based on ant colony optimization algorithm of swarm intelligence

Abstract

Traditional high-performance language parallel database processing learning has limitations in optimization algorithms, which is prone to falling into local optimal problems and has slow convergence. The pheromone update mechanism of the ant colony optimization algorithm lacks flexibility and is difficult to adapt to dynamic environments. This article improves the processing efficiency of high-performance language parallel database based on the improved ant colony optimization algorithm. The pheromone evaporation and reinforcement mechanism is applied in the algorithm, simulating the decay of pheromones in the natural environment and dynamically adjusting the pheromone concentration according to the path quality to guide ants to choose a better path. To cope with high-concurrency queries, a scheduling optimization strategy based on task priority is implemented, in which a weighted round-robin strategy is applied to dynamically allocate task resources to ensure that high-priority tasks can be processed in a timely manner. In addition, hash-based and range-based data partitioning methods are used to optimize data storage distribution. Through the application in high-performance language parallel database, the improved algorithm is compared and analyzed with the traditional algorithm. Under high concurrency pressure and large-data amount query load, the average query response time of the algorithm in this article is about 465.32 ms; the throughput is about 315.67 times per second; the CPU and memory resource utilization rates are 64.83% and 72.24%. The comparison results of t-test are p < 0.05, which shows that there are significant differences in the performance of the algorithm in this article compared to Dijkstra algorithm and genetic algorithm in various performance indicators. The research results indicate that the improved ant colony optimization algorithm has good application potential and can effectively improve the processing performance of parallel databases.