SOME SUFFICIENT CONDITIONS FOR OPTIMALITY IN CONTROL PROBLEMS WITH STATE SPACE CONSTRAINTS,
Abstract
The sufficient conditions obtained in the paper are an outgrowth of the work of Mangasarian. As in his paper a chain of inequalities and some ad hoc assumptions lead to a simple and direct proof of the main results. However the hypotheses are weaker, a somewhat different problem is treated, and jumps in the 'multipliers' corresponding to the differential constraints are allowed. The inclusion of the jumps is important for without them it is impossible to prove optimality in almost all optimal control problems where there is a state constraint of the form g(x,t)= or <0. If certain convexity and normality assumptions are imposed in the optimal control problem considered, the sufficient conditions become necessary. The sufficient conditions given here also apply to a number of interesting optimal control problems without state space constraints, including those where sufficient conditions of a similar type have been obtained previously. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1969
- Accession Number
- AD0706388
Entities
People
- E. G. Gilbert
- J. E. Funk
Organizations
- University of Michigan