Design and implementation of adaptive architecture for complex algorithms

作者：Ping Zhu, Weiming Zou, Pohua Lv, Jin Shi, Xuetao Jiang, Yirong Ma · 年份：2025 · DOI：10.1117/12.3044930 · 被引用次数：1 · 研究领域：Evolutionary Algorithms and Applications、Metaheuristic Optimization Algorithms Research、Neural Networks and Applications

Generally, large-scale or intelligent software development has the programming characteristics of complex algorithm, such as uncertain requirements, complex workflows, and frequent modifications. Under the constraint of limited perspective, studying progressive logic refinement and adaptive architecture have great significance for complex algorithm development. Starting from the analysis of the task characteristics, the big decision tree theory was proposed to solve complex algorithm development problems, and the methods for constructing, expanding, and running big decision trees were introduced. Focusing on the dynamic addition or modification of logic conditions, adaptive running mechanism, and running of big decision tree systems, this paper proposed the adaptive architecture and its implementation techniques. Taking the technique implementation of state monitoring and robust debugging as examples, the implementation methods of adaptive architecture were introduced. Finally, the future directions were summarized.