Intelligent evolution and algorithm realization path of decision support model in information system

Driven by the wave of digital transformation, information system has become the core support platform for enterprises and organizations to make decisions. However, the traditional decision support model has obvious limitations in the dynamic environment, and it is difficult to cope with the rapid changes of real-time data flow, user needs and business scenarios, as well as the high requirements of large-scale data and complex models for computational efficiency. In recent years, the breakthrough of AI technology provides a new path for the intelligent upgrade of decision support model. This paper aims to explore the intelligent evolution path of decision support model in information system by constructing a multi-level collaborative optimization framework. The dynamic evolution model based on markov decision processes (MDP) is studied, and the adaptive adjustment of the model is realized by combining online learning and meta-learning. In order to improve the efficiency and flexibility of decision support system (DSS), a hybrid optimization algorithm engine is developed and a hierarchical architecture supporting real-time decision is built. The experimental results show that this method is superior to the traditional DSS in different dynamic environments, especially in the severe disturbance environment, the accuracy is improved by 16.1%, showing stronger adaptability and stability. At the same time, this method is significantly better than the traditional method in response time, computing power consumption, memory occupation and energy consumption, showing higher computing efficiency and energy saving advantages. In addition, by introducing constraints based on SHAP value, the interpretability of the model evolution process is effectively guaranteed and the transparency of the model is ensured.