基于简约SQP和混合自动微分反应参数优化.pdfVIP

38 12 浙 江 大 学 学 报( ) V ol 38 N o 12 2004 12 J ournal of Zhejiang U niversity( Engineering Science) Dec 2004 SQP 江爱朋, 邵之江, 钱积新 ( , 3 10027) : , ( rSQP) , , , , , , ( SQ P) 100 , SQ P 10 , : ; SQ P; ; ; : T P202 7 : A : 100 8 97 3X( 2004) 12 1606 05 Optimization of reaction parameters based on rSQP and hybrid automatic differentiation algorithm JIAN G Aipeng, SHA O Zh ijiang, QIA N Jix in ( National Key Laboratory of Industrial Control Technology, Institute of Systems Engineering, Zhej iang University , angzhou, 310027, China) Abstract: T o opt imize t he react ion parameters in t he dy nam ical process of meth anoltohydrocarbons, a met hod b ased on reduced sequent ial quadrat ic program ming ( rSQP) and hybrid aut omat ic dif ferent iation t echnology w as present ed T he dynamical optimizat ion problem w as f irst ly discredit ed as a nonlinear programm ing problem denot ed by algebraic equat ion s With t he charact eristics of sparseness , relatively low degrees of freedom and equ ality constraints ut ilized , the nonlinear programming problem w as solved by improved rSQP solver In the solving process, hybrid aut omat ic diff erent iation t echnology w as u sed t o obt ain t he sparse stru cture and g radient inf orm at ion Comput at ional results show th at t he eff iciency of t he proposed method is more th an 100 t imes as t hat of t he st andard sequential quadratic programming ( SQP) w it h dif ferences, and is more t han 10 t imes as t hat of st andard SQP w it h hybrid autom at ic