Aiming at the problem that the performance of lightweight learning-based congestion control algorithms will fall off a cliff in some scenarios， a transmission control protocol congestion control switching scheme based on scenario change was proposed. Firstly， the real-time network environment was simulated by this scheme. Then， the scenario was identified according to the real-time environment parameters. Finally， the current congestion control algorithm was switched to the relatively optimal lightweight learning-based congestion control algorithm in this scenario. Experimental results prove that the proposed scheme is able to significantly improve network performance compared to the original schemes using a single congestion control algorithm， such as congestion control based on measuring Bottleneck Bandwidth and Round-trip propagation time （BBR） and Performance-oriented Congestion Control （PCC） with a total throughput increase of more than 5% and a total delay drop of more than 10%.

In order to solve the problems such as high time cost， inaccuracy and influence of parameter setting on algorithm performance when optimizing parameters of Convolutional Neural Network （CNN） by traditional manual methods， a variable Convolutional AutoEncoder （CAE） method based on Teaching-Learning-Based Optimization （TLBO） was proposed. In the algorithm， a variable-length individual encoding strategy was designed to quickly construct the CAE structure， and stack CAEs to a CNN. In addition， the excellent individual structure information was fully utilized to guide the algorithm to search the regions with more possibility， thereby improving the algorithm performance. Experimental results show that the classification accuracy of the proposed algorithm achieves 89.84% when solving medical image classification problems， which is higher than those of traditional CNN and similar neural networks. The proposed algorithm solves the medical image classification problems by optimizing the CAE structure and stacking CNN， and effectively improves the classification accuracy of medical image classification.

A multi-objective particle swarm optimization algorithm with Pareto neighborhood crossover operation (MPSOP) is proposed to solve the defect of local search for prrticle swarm optimization problems. MPSOP employs particle swarm optimization algorithm and Pareto neighborhood crossover operation to generate new population. A scaling factor used to balance contributions of particle swarm optimization algorithm and Pareto neighborhood crossover operation. Numerical experiments are compared with NSGA-II, SPEA2 and MOPSO on six benchmark problems. The numerical results show the effectiveness of the proposed MPSOP algorithm.

Traditional Mean-shift target tracking algorithm is rather sensitive to background environment. The use of kernel weighted histogram for target template and target candidates can not always get exact center of the target. Therefore, an improved feature selection mechanism was proposed, in which background weighted histogram was chosen for target template and kernel weighted histogram for target candidates. The simulation results show that the method achieves more accurate target tracking in complex environments.