We consider the control of the diserete component n_(t)of a owitching Markov proceaa x_(t)=(z_(t),n_(t))when there ia a running cost and an immediate coat c(i,j)for owitching n_(t)from i to j.We satudy the minimizatio...We consider the control of the diserete component n_(t)of a owitching Markov proceaa x_(t)=(z_(t),n_(t))when there ia a running cost and an immediate coat c(i,j)for owitching n_(t)from i to j.We satudy the minimization of the ergodic(or long-term average)total coat.Eooentially,this paper trento the cnce where,for n_(t)=n fixed,z_(t)ia a reflected diffusion or a reflected diffusion with jumps,nt being,for fixed z,a continuous-time Markov chain.Using the vanishing discount appronch,we exctend existing reoulta dealing with the situation where nt evolvea only by the switching control action and the diffusion is non-degenerate.Moreover,we solve the ergodic problem for a claso of diffusiono which can be degenerate and for an example with aboorbing atate.展开更多
文摘We consider the control of the diserete component n_(t)of a owitching Markov proceaa x_(t)=(z_(t),n_(t))when there ia a running cost and an immediate coat c(i,j)for owitching n_(t)from i to j.We satudy the minimization of the ergodic(or long-term average)total coat.Eooentially,this paper trento the cnce where,for n_(t)=n fixed,z_(t)ia a reflected diffusion or a reflected diffusion with jumps,nt being,for fixed z,a continuous-time Markov chain.Using the vanishing discount appronch,we exctend existing reoulta dealing with the situation where nt evolvea only by the switching control action and the diffusion is non-degenerate.Moreover,we solve the ergodic problem for a claso of diffusiono which can be degenerate and for an example with aboorbing atate.
