Computer Network Protocol Optimization in Computer Software Engineering

In dynamic 5G/6G and IoT environments, traditional TCP/IP protocols often experience reduced throughput and latency jitter. To tackle these issues, this study introduces protocol optimizations that incorporate an improved BIC-TCP with a dynamic adjustment factor, NSGA-II-based multi-objective parameter optimization, and DQN-driven dynamic strategies. The results demonstrate significant improvements, with the 99th percentile latency dropping to 179 ms at 90% load, representing a 45.4% reduction compared to traditional schemes. This optimized scheme offers robust performance, providing valuable support for the reconstruction of software-defined network protocol stacks.