1: \begin{abstract}

2: A standard approach to optimizing long-run running costs of discrete 

3: systems is based on minimizing the \emph{mean-payoff}, i.e., the long-run

4: average amount of resources (``energy'') consumed per transition. However, 

5: this approach inherently assumes that the energy source has an 

6: unbounded capacity, which is not always realistic. For example,

7: an autonomous robotic device has a battery of finite capacity that

8: has to be recharged periodically, and the total amount of energy consumed

9: between two successive charging cycles is bounded by the capacity.

10: Hence, a controller minimizing the mean-payoff must obey this 

11: restriction.

12: %

13: In this paper we study the controller synthesis problem for

14: \emph{consumption systems} with a finite battery capacity, where the

15: task of the controller is to minimize the mean-payoff while preserving

16: the functionality of the system encoded by a given linear-time

17: property. We show that an optimal controller always exists, and it may

18: either need only finite memory or require infinite memory (it

19: is decidable in polynomial time which of the two cases

20: holds). Further, we show how to compute an effective description of an

21: optimal controller in polynomial time.  Finally, we consider the limit

22: values achievable by larger and larger battery capacity, show that

23: these values are computable in polynomial time, and we also analyze

24: the corresponding rate of convergence.  To the best of our knowledge,

25: these are the first results about optimizing the long-run running

26: costs in systems with bounded energy stores.

27: \end{abstract}

28: