基于市场机制的QoS控制模型MQC的改进遗传算法求解

基于市场机制的QoS控制模型(MQC)的求解是一个NP-Hard问题，对此作者提出一种改进遗传算法(MGA)用于求解，与传统遗传算法相比加入自适应预测器，同时在遗传操作中采用自适应遗传因子指导搜索过程．最后通过实验证明该算法快速有效。

Solving Market-Based QoS Control Model with Improved Genetic Algorithm

This paper proposes an improved genetic algorithm (MBA) to solve the market-based QoS control model (MQC) which uses market mechanisms to provide QoS control services and essentially a NP-Hard problem. MQC is carefully studied, two key components of MBA are introduced: a QoS prediction strategy and an adaptive genetic algorithm. The QoS prediction strategy predicts the QoS parameters for future using adaptive linear prediction that minimizes the mean square error, and so comes the name Self-Adaptive Least Mean Square Line Predictor (SLMSLP). The adaptive technique of SLMSLP does not require any prior knowledge of the QoS statistics, nor assume stationarity. The results for QoS prediction from an experimental system verifies the better prediction capability of SLMSLP under different parameters and numbers of application. Besides with SLMSLP, an adaptive genetic algorithm is proposed to compute the best price and network resource distribution in MQC. Contrasted with traditional genetic algorithm, the adaptive genetic algorithm has an exchange of crossover and mutation in sequence and adaptive probabilities of crossover and mutation, which protect the diversity of colony. The performance comparison from an experimental system shows that the MBA algorithm is effective and efficient with computing time reduced by over 20%.