A boundary-value technique to solve linear state regular problems (Q1114643)

A method is proposed to solve free endpoint optimal control problems with linear state and quadratic cost. After translating the problem into a two-point boundary-value problem (TPBVP) that arises from the optimization of the Hamiltonian, two points \(t_ 1\) and \(t_ 2\), \(t_ 1<t_ 2\), belonging to the given interval of integration \([t_ 0,t_ f]\), are chosen and conditions are derived at these points appropriately. Then, letting \(\tau =(t-t_ 0)/\epsilon\), \(\sigma =(t_ f-t)/\epsilon\), the \(\tau\)-scaled, the original, and the \(\sigma\)-scaled TPBVP's are solved on the intervals \([t_ 0,t_ 1]\), \([t_ 1,t_ 2]\), and \([t_ 2,t_ f]\) respectively as boundary-value problems.