An error correction algorithm for stochastic production costing

A theoretical approach is presented for calculating the error involved in the solution of stochastic production costing problems using the method of cumulants. It is proposed to use this calculated error in an algorithm for obtaining an improved stochastic approximation of the equivalent-load-duration curve. The results show that of the three series that can be used for stochastic approximation, namely, the beta, the Gram-Charlier, and the Laguerre, the beta gives the best results. All three methods, however, give unacceptable errors in the approximation of the expected unserved energy. Another method is proposed that approximates the error between the beta curve and the actual curve using the Fourier-Chebyshev expansion, which is used to develop a better stochastic approximation. Results obtained using practical system data indicate that this method gives much better results than the three methods discussed above