Criteria for the well-posedness of a linear two-point boundary value problem for systems of integro-differential equations (Q610321): Difference between revisions

The authors study a linear two-point boundary value problem for systems of Fredholm integro-differential equations \[ {dx\over dt}= A(t)x+ \int^T_0 K(t,s)x(s)\,ds+ f(t),\quad T> 0,\;t\in [0,T],\;x\in \mathbb{R}^n,\tag{1} \] \[ Bx(0)= Cx(T)= d,\quad d\in\mathbb{R}^n,\tag{2} \] where the \(n\times n\) matrices \(A(t)\) and \(K(t,s)\) are continuous on \([0,T]\), respectively \([0,T]\times [0,T]\), \(B\) and \(C\) are constant \(n\times n\) matrices, \(\| x\|= \max_{1\leq i\leq n}|x_i|\) for \(x= (x_1,\dots, x_n)\), \(\| A(t)\|= \max_{1\leq i\leq n}\, \sum^n_{j= 1}|a_{ij}(t)|\leq \alpha\) for \(A(t)= (a_{ij}(t))_{ij}\) and \((\forall(t,s)\in [0,T]\times [0,T])(\| K(t,s)\|\leq\beta\)) for \(\alpha> 0\), \(\beta> 0\) constants. The authors divide the interval \([0,T]\) into parts with increment \(h> 0\), \(mh= T\) and approximate (1), (2) by the two-point boundary value problem for the systems of loaded differential equations: \[ {dy\over dt}= A(t)y+\sum^m_{i=1}\;\int^{ih}_{(i-1)h} K(t,s) y((i- 1)h)\,ds+ f(t),\;t\in [0,T],\;y\in\mathbb{R}^n,\tag{3} \] \[ By(0)+ Cy(T)= d,\quad d\in \mathbb{R}^n.\tag{4} \] The boundary value problem (1), (2), respectively (3), (4), is said to be well-posed if there exist \(\gamma> 0\) and \(K> 0\), such that for arbitrary \(f\) and \(d\), there exist a unique solution \(x^*\), respectively \(y^*\), and the estimate \[ \| x^*\|_1= \max_{t\in [0,T]}\,\| x^*(t)\|\leq\gamma\, \max(\| f\|_1,\| d\|), \] respectively \[ \| y^*\|_1\leq K\,\max(\| f\|_1,\| d\|). \] Two main results proved by the authors affirm that if denote \[ \delta_1= ((\alpha+ \beta T)+1)\beta Th,\quad \delta_2= ((\alpha+\beta T)K+1)\beta Th, \] then: (i) if the problem (1), (2) is well-posed with constant \(\gamma\) and \(q_1(h)= 2^{-1}\delta_1< 1\), the boundary value problem (3), (4) is well-posed with constant \(K= 2\gamma(2- \delta_1)^{-1}\) and \[ \| y-x\|_1\leq \gamma\delta_1(2- \delta_1)^{-1}\,\max(\| f\|_1,\| d\|), \] where \(x\), respectively \(y\), are the solutions of problem (1), (2), respectively (3), (4), (ii) if the problem (3), (4) is well-posed with constant \(K\) and \(q_2(h)= 2^{-1}\delta_2< 1\), the boundary value problem (1), (2) is well-posed with constant \(\gamma= 2K(2-\delta_2)^{-1}\) and \[ \| x-y\|_1\leq K\delta_2(2- \delta_2)^{-1}\,\max(\| f\|_1,\| d\|), \] where \(x\), respectively \(y\), are solutions of problem (1), (2), respectively (3), (4). The authors study the problem (3), (4) by a parametrization method.