Re: Re: Re: ricatti & set of ODE solution.
- To: mathgroup at smc.vnet.net
- Subject: [mg41737] Re: [mg41716] Re: [mg41692] Re: [mg41676] ricatti & set of ODE solution.
- From: Arda Kutlu <e130559 at metu.edu.tr>
- Date: Tue, 3 Jun 2003 07:13:19 -0400 (EDT)
- Sender: owner-wri-mathgroup at wolfram.com
> I'm still not understanding the nature of your system. It sounds as if > you have a system of 1st order differential equations with initial > values at t=0 (?) and final values at t=2. That kind of problem is > simply not well posed. Or do you have second order equations? allright. here is the question. ricatti part only. R=r*IdentityMatrix[4] r is a value to be found or chosen. Q=q*IdentityMatrix[4] q is also a value to be chosen. ricatti equation. M(K) --> matrix of K 4x4 Mprime(K) is the time derivative of K. Mprime(K)= -M(K).Transpose[A].M(K)-Q+M(K).B.R.Transpose[B].M(K) where A={{0,1,0,0}, {0,0,0,0}, {0,0,0,1}, {0,0,0,0}} B={{0,0}, {0.116, (-0.116)}, {0, 0}, {(-0.116), 1.366}} first order nonlinear differential set. and boundary . K(2)=H value of K at 2 seconds when the proceses ends. H=h*IdentityMatrix[4] again h is a value to be chosen. As you can see here i only know/chose the final value of K. And I need to find K. Possible numerical solution to this problem is shooting method, but due to the last term in ricatti the ode is non-linear and the method doesn't work. I am supposed to find values of K in 0 2 seconds interval with a small increment. Or better K(t). it is well possed. Just the problem is i don't know the initial values but final values. These were for ricatti solution. If u are interested in optimization or this problem i can send you the problem and my solution (my solution using the other way - set of odes -with semi-numerical solution, symbolic solution looks like not possible with today's pc).