Computer Science > Distributed, Parallel, and Cluster Computing
arXiv:2511.01255 (cs)
Abstract:This paper focuses on the key problem in the development of nonlinear optical technology, the performance optimization of aperiodically polarized crystals. The performance of the crystal depends on the precise control of the micro distribution of crystal domains, but its optimization belongs to the high-dimensional discrete combination “NP hard” problem. The traditional algorithm has the bottleneck of slow convergence and easy to fall into local optimization, while the heuristic methods such as genetic algorithm are limited by the CPU serial calculation and inefficient. In order to solve the above challenges, this paper proposes the …
Computer Science > Distributed, Parallel, and Cluster Computing
arXiv:2511.01255 (cs)
Abstract:This paper focuses on the key problem in the development of nonlinear optical technology, the performance optimization of aperiodically polarized crystals. The performance of the crystal depends on the precise control of the micro distribution of crystal domains, but its optimization belongs to the high-dimensional discrete combination “NP hard” problem. The traditional algorithm has the bottleneck of slow convergence and easy to fall into local optimization, while the heuristic methods such as genetic algorithm are limited by the CPU serial calculation and inefficient. In order to solve the above challenges, this paper proposes the fusion scheme of hwsda hybrid optimization algorithm and GPU parallel acceleration technology: the differential evolution algorithm (DE) is used to realize the global search, and the gray wolf optimization algorithm (GWO) is used to strengthen the local search and convergence speed, and the two coordinate to balance the global and local optimization requirements; At the same time, it relies on GPU multi-core architecture to realize thread level parallel computing and improve optimization efficiency. This scheme effectively breaks through the optimization problem of high-dimensional discrete space, improves the accuracy of crystal domain control, improves the efficiency of quasi phase matching design by hundreds to thousands of times compared with traditional CPU serial computing, provides a new paradigm for the design of complex nonlinear optical devices, and helps promote the performance breakthrough and industrial application of related devices in the fields of quantum optics and laser processing.
| Subjects: | Distributed, Parallel, and Cluster Computing (cs.DC) |
| Cite as: | arXiv:2511.01255 [cs.DC] |
| (or arXiv:2511.01255v1 [cs.DC] for this version) | |
| https://doi.org/10.48550/arXiv.2511.01255 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Chen He [view email] [v1] Mon, 3 Nov 2025 05:56:02 UTC (1,451 KB)