The following steps are to be followed meticulously in order to develop a very reliable and valid model for the energy industry:

• Define the expected outcome/objective of the model

• Outline the general process and data collection

• Identify the manipulable and nonmanipulable variables by observed data, analytical data, or a combination of both

• Develop a basic model with the identified variables which affect the objective function and deleting the least effective variables by model revalidation

• Perform sensitivity analysis and model validation by manipulating the variables and comparing the observed output with model output

• Check model validity under observed conditions for a stipulated range of variables

• No validation will be deemed necessary if the deviation between observed and model outputs are within statistical limits

Once the last stage has been reached, the model will be accepted as valid and will form a basic DSS tool, by which the actual performance of the system may be evaluated and compared and the reason for any discrepancy may be identified for remedial action.

If the variation is explainable, the decision-maker may correct the situation immediately. If the variation is not explainable or is exceeds the stipulated limits, the situation must be taken as out of control and warrants thorough investigation. Hence, EORT models are capable of giving the warning signals at the right time and can easily avert inefficiency or even disaster in some cases.