Math Meth Oper Res
DOI 10.1007/s00186-015-0496-y
ORIGINAL ARTICLE
Continuity of the optimal average cost in Markov
decision chains with small risk-sensitivity
Selene Chávez-Rodríguez · Rolando Cavazos-Cadena ·
Hugo Cruz-Suárez
Received: 5 October 2014 / Accepted: 10 February 2015
© Springer-Verlag Berlin Heidelberg 2015
Abstract This note concerns discrete-time controlled Markov chains driven by a
decision maker with constant risk-sensitivity λ. Assuming that the system evolves on
a denumerable state space and is endowed with a bounded cost function, the paper
analyzes the continuity of the optimal average cost with respect to the risk-sensitivity
parameter, a property that is promptly seen to be valid at each no-null value of λ.
Under standard continuity-compactness conditions, it is shown that a general form of
the simultaneous Doeblin condition allows to establish the continuity of the optimal
average cost at λ = 0, and explicit examples are given to show that, even if every state
is positive recurrent under the action of any stationary policy, the above continuity
conclusion can not be ensured under weaker recurrence requirements, as the Lyapunov
function condition.
Keywords Risk-sensitive optimality equation · Discounted approach · Uniformly
bounded solutions to the optimality equation · Jensen’s inequality · Stopping times ·
Transient states · Continuity at risk-neutrality
This work was supported by PSFO under Grant No. 14-300-01, and by PRODEP under Grant
No. 17332-UAAAN-CA-23.
S. Chávez-Rodríguez · H. Cruz-Suárez
Facultad de Ciencias Físico-Matemáticas, Ave. San Claudio y Río Verde, Col. San Manuel CU,
Benemérita Universidad Autónoma de Puebla, 72570 Puebla, PUE, Mexico
e-mail: selenechavez@alumnos.fcfm.buap.mx
H. Cruz-Suárez
e-mail: hcs@fcfm.buap.mx
R. Cavazos-Cadena (B )
Departamento de Estadística y Cálculo, Universidad Autónoma Agraria Antonio Narro,
Boulevard Antonio Narro 1923, Buenavista, 25315 Saltillo, COAH, Mexico
e-mail: rcavazos@uaaan.mx
123