Optimal incentive control for two-level dynamical constrained problems: a solution method through sensitivity analysis
P. Cornet, M. Installé · 2002
Incentive control of hierarchical bilevel planning problems cannot be seen as usual optimal control problems since the upper-level decision maker, who wants to control the dynamical evolution of the system, has no direct means of action on the system's decision variables, which are in the hands of the lower-level decision maker. However, he can ultimately trigger changes in the lower-level strategies through indirect actions called incentives. This two-level interaction between actors, with different control variables and different objective functions, can be seen as a dynamical constrained Stackelberg game. A thorough description and formulation of the problem characterized by a multi-objective dynamical constrained optimization and a hierarchical framework is given. An algorithm intending to solve this problem using sensitivity analysis, and an example from a regional planning context illustrating the concept of control for the upper-level actor are included.>