scientific article; zbMATH DE number 3907224

From MaRDI portal
Publication:3684718





The following time-invariant system is given: \(\dot x=Ax+bu,\) \(x(0)=x_ 0\neq 0,\) \(x(t_ 1)=0,\) \(x\in R^ n,\) \(u\in R^ 1,\) \(b\in R^ n,\) where \(\dim A=n\times n,\) \(t_ 1=\min \{t\}.\) Using the maximum principle, a time-optimal control law is derived in the form: \((*)\quad u(t,p)=-sign[p_ 0\gamma (t)],\) where \(p_ 0=p(0,p_ 0)\) is the initial data of the conjugate vector \(p(t,p_ 0)\), \(\gamma (t)=\exp (- tA)b,\) exp(-tA) is a fundamental matrix. The aim of the paper is to find initial data \(p_ 0=p_ 0(x_ 0)\) in order to define the optimal control u(t,p) on \([0,t_ 1]\) of the 'bang-bang' type (*). - One obtains the initial data \(p_ 0\) after solving the following finite-measure system: \[ \xi (p,t_ 1)=x_ 0,\quad | p|^ 2=1,\quad where\quad \xi (p,t_ 1)=-\int^{t_ 1}_{0}\exp (-tA)b sgn[p\gamma (t)]dt. \] For the solution of that system a recursive method is suggested. This method, however, gives real results only for simple models of the control-object, thus it has a very small practical significance. The reviewer recommends the method of initial values of conjugate state search, proposed in the works by \textit{J. H. Eaton}, J. Math. Anal. Appl. 5, 329-344 (1962; Zbl 0142.068)] and \textit{W. G. Boltyanskij} [Mathematical methods in optimal control (1966; Zbl 0198.199)]. This method is useful in a large area of practical problems.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3684718)