Multi-Cyclic Queuing and Forwarding (Multi-CQF) extends conventional Cyclic Queuing and Forwarding (CQF) by introducing multiple queue groups with different cycle durations, enabling efficient support for heterogeneous traffic flows in Time-Sensitive Networking (TSN). However, Multi-CQF scheduling remains a challenging optimization problem due to the joint consideration of flow mapping, cycle constraints, bandwidth allocation, and time injection scheduling. This paper proposes an Improved Genetic Algorithm with Incremental Queue Reassignment (IGA-IQR) for Multi-CQF scheduling. First, a score-guided flow sorting strategy is introduced to prioritize flows according to deadline and packet-size characteristics. Second, a constraint-oriented fitness function is designed to reduce deadline and bandwidth violations during genetic optimization. Third, an Incremental Queue Reassignment (IQR) mechanism is proposed to locally repair unschedulable flows with low computational overhead. Experimental results under different network loads, topologies, and traffic specifications demonstrate that the proposed method achieves higher scheduling success rates and better robustness compared with the baseline GA approach. The results show that IGA-IQR provides an effective and scalable solution for Multi-CQF scheduling in heterogeneous TSN environments.

Time-Sensitive Networking (TSN), Cyclic Queuing and Forwarding (CQF), Multi-CQF, Genetic Algorithm, Traffic Scheduling.