基于智能优化CNN模型的高速公路短时交通流量预测

打开文本图片集

Short-Term Traffic Flow Prediction of Expressways Based on an Intelligently Optimized CNN Model

CHENJiahang，Hlinngin（olgringanUitysan）

关键词：高速公路；交通流量预测；变分模态分解；CGO-CNN模型；深度学习中图分类号：F259.27；U231 文献标志码：AD0I：10.13714/j.cnki.1002-3100.2026.04.029

Abstract：Theresearchonthe predictionof short-term traic flow inthe expressway environmentcanplayanimportant role in alleviating congestionandoptimizing trafic management.This paper proposesa hybrid modelbasedon Chaos Game Optimization （CGO）and Convolutional Neural Networks （CNN）deep learning，referred toas the CGO-CNN model. The model utilizes VariationalModeDecomposition（VMD）todenoisetraicatatherebyiprovingdataquality.Iteffectivelyintegratestheglobal searchcapabilityofCGOandthenonlinearmappngabilityofCNN，enabling moreaccuratetraficflowpredictions.Experimental results show thattheCGO-CNNmodeloutperforms traditionalmodels invarious metricsespeciallyintermsofMAPEandRMSE. This demonstrates the model's application potential and reliability in intelligent transportation systems.

Key Words： expressway;trafic flow prediction; Variational Mode Decomposition; CGO-CNN Model; deep learning

0引言

伴随城市化进程加速推进及机动车保有量持续增长，交通拥堵和环境污染等问题，因严重影响居民生活质量，已成为公众关注的焦点。（剩余4259字）