Pathways to Hopf bifurcations in dynamic continuous-time optimization problems (Q1359445)

The paper considers a class of optimal control problems in continuous time, having a scalar control variable, two state variables (with an initial condition), an infinite time horizon, and a Lagrange-type objective function. [To give the model some economic flavour, a discount factor is introduced in the objective.] An important assumption is that the Hamiltonian is supposed to be concave in control and states, and that a limiting transversality condition is satisfied. The purpose of the paper is to identify conditions under which stable limit cycles can appear as optimal solutions of such control problems. The paper modifies and extends previous results, obtained in [\textit{F. Wirl}, Ann. Oper. Res. 37, No. 1-4, 345-356 (1992; Zbl 0784.90024)]. It is shown that there are (at least) two possible ways to obtain stable limit cycles as optimal solutions in the model class under consideration: state-control interactions in the model structure, and growth conditions. Economic examples demonstrating the application of these two methods are given.