Forecasting concepts and Bayesian decision theory applied to production control problems

Methods for formulating decision rule 5 are demonstrated for two block control production problems. Exponential smoothing techniques are applied to a sequence of the number defective units observed in successive batches and are used to forecast the future of the production process. The forecast statistics are used to quantify the decision maker's prior beliefs about the outcome of a future observation. Bayesian decision theory is employed to compare the consequences of a set of alternative actions and to select the best decision among those available. The particular production control problems utilize the general form the quadratic exponential smoothing model and employ a techniq1 for revising the forecasts and forecast statistics to -include sample evidence.