模糊物元分析在经验库算法中的性能优化

经验往往是决定任务成败的关键因素。算法在原有经验库算法基础之上引入特征因子和模糊物元分析过程，首先对发生事件或任务作归一化代码处理，接着搜索存在的解决方案及组合解决方案并将其存入有序解集，然后将有序解集中各解决方案的特征因子作为输入进行处理，形成最优解决方案，最后更新有序解集中各解决方案的动态特征因子。实验表明，优化算法增加了处理时间，但有效提升了解的优劣度，改进了经验库的使用性能，更好地满足了实际的需要。

Performance Optimization of Experience-Database Based on Fuzzy Matter-Element Analysis

A fractionation protocol of soluble COD (SCOD) was put forward by combining respirometric method with hydrolysis kinetics of soluble slowly biodegradable COD (S_H). SCOD fractionation results of two wastewater treatment plants (WWTP) in Shanghai show that the SCOD in sand basin effluents (typical domestic wastewater) of WWTP A is composed of 43.5%-58.6% S_H, 21.8%-35.2% readily biodegradable COD (S_S) and 15.4%-30.9% soluble inert COD (S_I), and those SCOD fractions in sand basin effluents (combined sewers after long pipeline transportation) of WWTP B are 34.5%-45.2%, 29.3%-37.7% and 25.6%-31.2%, respectively. The linear regression results of respirometric tests data from nine samples demonstrate that the first-order kinetics can reliably describe hydrolysis process of S_H, and the kinetic constants of S_H from WWTP A and B are respectively 28.00-39.77 d^-1 and 26.48-29.52 d^-1. Experimental results demonstrate that this protocol can achieve theoretical partition for the integration area of S_S, and also eliminate the effect of soluble microbial products on S_I determination.