In this paper, we discuss the problem of scheduling n jobs on a batch processor to minimize total earliness and tardiness. We propose a dynamic programming algorithm in polynomial time to show the problem is polynomially solvable. However, the algorithm is not efficient in practice because of the high exponent and large memory requirement. We then propose a dynamic programming algorithm to solve the problem in pseudopolynomial time. We also extend our analysis to the cases of weighted earliness and tardiness problems, jobs with agreeable release times, and jobs with due windows. Finally, we report and discuss computational results.
