System and method for tuning a segmented model representating product flow through a supply chain or manufacturing process
View Patent ↗A method of systematically applying parameter evaluation techniques to large numbers of complex segmented models used for managing and analyzing a segmented manufacturing or supply chain process. These models include a plurality of segment models, each segment model representing a process segment within the segmented manufacturing or supply chain process and comprising a mathematical expression including at least one parameter. The method compares predicted values generated by a segment model with actual observed values to determining a score representative of the performance of the segment model. New parameters for the mathematical equation associated with the segment model are determined through statistical techniques when the model score indicates an unacceptable performance of the segment model. The segment models may be evaluated and updated in parallel and without manual intervention.
1. A computer-implemented method for managing and modeling a segmented logistical process; the steps comprising:
creating a computer model of said segmented logistical process with a plurality of segment models, each segment model representing a process segment within said segmented logistical process and comprising a mathematical expression including at least one parameter; and
for each segment model:
acquiring actual observed values associated with the process segment corresponding to said segment model;
electronically comparing, by a computer, predicted values generated by the segment model with said actual observed values;
electronically determining, by said computer, a score representative of the performance of said segment model from said comparison of the predicted values with said actual observed values;
statistically determining, by said computer, new parameters for the mathematical equation associated with said segment model when said score indicates an unacceptable performance of said segment model; and
electronically updating, by said computer, said segment model with said new parameters.
2. The method for managing and modeling a segmented logistical process in accordance with claim 1 , wherein said step of updating said segment model with said new parameters includes the steps of:
electronically determining, by said computer, a score representative of the performance of said segment model utilizing said new parameters; and
electronically updating, by said computer, said segment model with said new parameters when said last-recited score indicates an acceptable performance of said segment model.
3. The method for managing and modeling a segmented logistical process in accordance with claim 1 , wherein said logistical process comprises a manufacturing process.
4. The method for managing and modeling a segmented logistical process in accordance with claim 1 , wherein said logistical process comprises a supply chain process.
5. A computer-implemented method for updating a segment model which represents a process segment within a segmented logistical process, said segment model comprising a mathematical expression including at least one parameter; the method comprising the steps of:
acquiring actual observed values associated with said process segment;
electronically comparing, by a computer, predicted values generated by the segment model with said actual observed values;
electronically determining, by said computer, a score representative of the performance of said segment model from said comparison of the predicted values with said actual observed values;
statistically determining, by said computer, new parameters for the mathematical equation associated with said segment model when said score indicates an unacceptable performance of said segment model; and
electronically updating, by said computer, said segment model with said new parameters.
6. The method in accordance with claim 5 , wherein said step of updating said segment model with said new parameters includes the steps of:
electronically determining, by said computer, a score representative of the performance of said segment model utilizing said new parameters; and
eletronically updating, by said computer, said segment model with said new parameters when said last-recited score indicates an acceptable performance of said segment model.
7. The method in accordance with claim 5 , wherein said logistical process comprises a manufacturing process.
8. The method in accordance with claim 5 , wherein said logistical process comprises a supply chain process.
9. In a system for modeling a segmented logistical process with a plurality of segment models, each segment model representing a process segment within said segmented logistical process, and each segment model comprising a mathematical expression including at least one parameter; the improvement comprising:
means for updating said segment models, said means performing the steps of:
acquiring actual observed values associated with a process segment;
comparing predicted values generated by the segment model representing said process segment with said actual observed values;
determining a score representative of the performance of said segment model from said comparison of the predicted values with said actual observed values;
statistically determining new parameters for the mathematical equation associated with said segment model when said score indicates an unacceptable performance of said segment model; and
updating said segment model with said new parameters.
10. The system in accordance with claim 9 , wherein said logistical process comprises a manufacturing process.
11. The system in accordance with claim 9 , wherein said logistical process comprises a supply chain process.