Why (some) abnormal problems are ``normal'' (Q2242910)

The authors discuss the question of `abnormality' of certain infinite time horizon problems going back to [\textit{H. Halkin}, Econometrica 42, 267--272 (1974; Zbl 0301.90009)]. Here, abnormality means a lack of constraint qualifications and thus only optimality conditions of John-type holding for the solution. The authors show that the abnormality is generic in the sense that for a particular model only degenerate Lagrange-multipliers can be obtained if the initial values are too large compared to the discount rate in the infinite horizon objective.