This paper proposes the application of vectorization and parallelizati
on techniques for solving the unit commitment problem on supercomputer
s. A modified dynamic programming method which takes full advantage of
the vector processors present in a supercomputer is presented. Parall
el processing is used to further speedup the computation time. Compute
r simulation of a 26-unit power system is performed on a CRAY Y-MP2/21
6 in both scalar and vector codes. The computation time results for th
e conventional dynamic programming algorithm and modified algorithm ar
e compared to demonstrate the effectiveness of the proposed method.