Experiments for both parallel neural network and structure network with RLS algorithm had been done for simulated model with concentration reactant unmeasured.
英
美
- 作者使用并聯(lián)混合網(wǎng)絡(luò )和結構逼近式混合網(wǎng)絡(luò )分別為濃度不可測的CSTR仿真對象進(jìn)行建模，并對仿真結果進(jìn)行了比較，實(shí)驗結果表明利用結構逼近式混合網(wǎng)絡(luò )能夠以較高的精度為部分狀態(tài)可測的非線(xiàn)性動(dòng)態(tài)系統建模。
