Assume that the third period $(Wednesday)$ requirement $is$ not known for certain. $It$ could $be$ either $3000or9000.$ This won't $be$ known until after Tuesday's transaction has been made. Therefore the withdrawal schedule can $be$ changed for Wednesday through Sunday, after Tuesday, and upon learning the real demand for Wednesday.

The primary question $is$ what $is$ the "best decision" $in$ period $1-$ what should the withdrawal $beon$ Monday? And what kind $of$ errors are involved with making the wrong assumption about the demand for cash $on$ Wednesday. $If$ you make the wrong decision, how much could this "mistake" cost you? This $is$ not simply a question $of$ running $an$ optimization $of$ the two situations.

Determine the best withdrawal schedule $-$ which can include a change $in$ the original plan. Determine also the percent error $of$ making a decision based $on$ one $of$ the two possible options for Wednesday's demand.

Hint: Notice that the objective function value $is$ different for the two scenario optimal solutions, $so$ determining the error $is$ based $on$ the two scenarios.