Definition
g ( x ) : = { f s + c s x , b s − 1 ≤ x ≤ b s , ∀ s ∈ { 1 , 2 , … , S } g(x):=\bigg\{f^s+c^sx,\quad b^{s-1}\le x\le b^{s},\quad \forall s\in \{1,2,\dots,S\} g(x):={ fs+csx,bs−1≤x≤bs,∀s∈{ 1,2,…,S}
- c s c^s cs the slope
- f s f^s fs the cost-intercept
- b s − 1 , b s b^{s-1}, b^s bs−1,bs the breakpoints (upper and lower bounds). b 0 = 0 b^0=0 b0=0
Three Models for Piecewise Linear Costs
Define f ^ s = g s ( b s − 1 ) − g s − 1 ( b s − 1 ) = ( f s + c s b s − 1 ) − ( f s − 1 + c s − 1 b s − 1 ) \hat{f}^s=g^s(b^{s-1})-g^{s-1}(b^{s-1})=(f^s+c^sb^{s-1})-(f^{s-1}+c^{s-1}b^{s-1}) f^