A note on risk averse newsboy problem

This paper considers the maximization of expected utility functions in the classical newsboy problem. First, the general theory for the risk averse utility functions is developed. Then, the necessity of approximation occurs when the logarithmic utility, which is familiar to the risk theory, is used in the newsboy problem. Applying the Taylor approximation, the authors consistently define the logarithmic utility in the context of the newsboy problem. Secondly, they discuss an alternative problem maximizing the upper and lower bounds of the expected utility function, instead of the direct maximization of it. Some numerical examples show that the approximation procedure and the alternative objective proposed here are useful from the practical viewpoint.