You want to express the cost of the transaction as a function of the number of tables purchased by, say C(t), where t denotes the # of tables purchased and C(t) the cost thereof. We need to determine C(t) explicitly that we may deduce its extrema.
"Your firm offers to deliver 300 tables to a dealer, at $90 per table, and to reduce the price per table on the entire order by 25cent for each additional table over 300."
The easiest way to model the cost function is piecewise:
If 300 or less tables are sold, then the price of each table is $90, so
[tex]C(t)=90t,\mbox{ for } t\leq 300[/tex]
after that, they get cheaper, so we modify the cost function to reflect that change in price per table. For each table sold over 300, the price per table drops $0.25 from the $90 base price per table, that is [$90 - (discount=$0.25 per table)(# of times to apply discount=t-300)] is the price per table if they bought t tables and t>300, then multiply by the # of tables purchased to get the total cost. Hence
[tex]C(t)=\left[ 90-0.25\left( t-300\right) \right] t,\mbox{ for } t>300[/tex]
this can be simplified to [itex]C(t)=165t-0.25t^{2},\mbox{ for } t>300[/itex]
So putting it all togeather, we have the piecewise cost function C(t) given by
[tex]C(t)=\left\{\begin{array}{cc}90t,&\mbox{ if }<br />
t\leq 300\\165t-0.25t^{2}, & \mbox{ if } t>300\end{array}\right.[/tex]
From here, do what you would expect to: find the relative extrema of C(t).
Here is a plot of C(t) I did with maple...