Methods for supply chain management incorporating uncertainty

A robust method for solving in a programmed computer, optimization problems under uncertainty applicable to Analytical Hierarchical Processing (AHP) that enables a robust representation of the optimization problem, each optimization problem having (i) one or more objective functions, each of the objective functions having one or more constraints, wherein the number of constraints is typically smaller than the number of parameters, with one or more classes of uncertainty, (ii) input data represented in one or more ensembles with inputs being specified in one or more stages (iii) parameters based on the ensembles, (iv) one or more assumptions about a future, and (v) quantitative estimates of information within the input data and one or more outputs to obtain one or more ranking given one or more options depending on one or more qualitative or quantitative criteria, information being delineated into a hierarchy of criteria and alternatives comprising the steps of:

• specifying the uncertainty as a hierarchical series of sets of constraints on parameters, with the parameters restricted to each constraint set forming an ensemble, and the hierarchy of constraints, represented as mathematical sets forming a hierarchy of ensembles, said hierarchy being based on subset, intersection or disjoint relationships amongst them, the hierarchical series of sets of constraints on parameters represented in a storage element;
• utilizing a processor executing optimization techniques to create effective minimum and maximum bounds on each objective function, said bounds depending on the constraints in each ensemble of parameters and being generated responsive to each assumption about the future, the effective minimum and maximum bounds on each objective function represented in the storage element;
• estimating, with the processor, a volume of candidate ensembles and relating the volume to one or more information theoretic measures;
• utilizing the one or more information theoretic measures to analyze and improve the ensembles and associated stages by changing a specificity of the bounds on the objective function, the ensembles and information theoretic measures represented in a storage element;
• obtaining relative ratios and alternatives for each criteria;
• and
• generating a random matrix to satisfy a constraint set.

A robust method for solving in a programmed computer, optimization problems under uncertainty applicable to Auctions, which rely on a valuation benefit function that enables a robust representation of the optimization problem, each optimization problem having (i) one or more objective functions, each of the objective functions having one or more constraints, wherein the number of non-positivity constraints is typically smaller than the number of parameters, with one or more classes of uncertainty, (ii) input data represented in one or more ensembles with inputs being specified in one or more stages (iii) parameters based on the ensembles, (iv) one or more assumptions about a future, and (v) quantitative estimates of information within the input data to measure the benefit derived by acquiring a good of some kind, the method comprising:

• specifying the uncertainty as a hierarchical series of sets of constraints on parameters, with the parameters restricted to each constraint set forming an ensemble, and the hierarchy of constraints, represented as mathematical sets forming a hierarchy of ensembles, said hierarchy being based on subset, intersection or disjoint relationships amongst them, the hierarchical series of sets of constraints on parameters represented in a storage element;
• utilizing a processor executing optimization techniques to create effective minimum and maximum bounds on each objective function, said bounds depending on the constraints in each ensemble of parameters and being generated responsive to each assumption about the future;
• estimating, with the processor, a volume of candidate ensembles and relating the volume to one or more information theoretic measures;
• and
• utilizing the one or more information theoretic measures to analyze and improve the ensembles and associated stages by changing a specificity of the bounds on the objective function.

A robust method for solving in a programmed computer, optimization problems under uncertainty applicable to Analytical Hierarchical Processing (AHP) that enables a robust representation of the optimization problem, each optimization having (i) one or more objective functions, each of the objective functions having one or more constraints, wherein the number of constraints is typically smaller than the number of parameters, with one or more classes of uncertainty, (ii) input data represented in one or more ensembles with inputs being specified in one or more stages (iii) parameters based on the ensembles, (iv) one or more assumptions about a future, and (v) quantitative estimates of information within the input data and one or more outputs to obtain one or more ranking given one or more options depending on one or more qualitative or quantitative criteria, information being delineated into a hierarchy of criteria and alternatives further comprising the steps of:

• specifying the uncertainty as a hierarchical series of sets of constraints on parameters, with the parameters restricted to each constraint set forming an ensemble, and the hierarchy of constraints, represented as mathematical sets forming a hierarchy of ensembles, said hierarchy based on subset, intersection or disjoint relationships amongst them, the hierarchical series of sets of constraints on parameters represented in a storage element;
• utilizing optimization techniques in the programmed computer to identify minimum and maximum bounds on each objective function, said bounds depending on the constraints in each ensemble of parameters and computed for each of the assumptions;
• estimating a volume of candidate ensembles and relating the volume to one or more information theoretic measures;
• utilizing the one or more information theoretic measures to analyze and iteratively refine the ensembles with the programmed computer by changing a specificity of said bounds on the objective function;
• obtaining relative ratios and alternatives for each criteria;
• and
• generating a random matrix with the programmed computer to satisfy a constraint set.

A computer program product embodied in a non-transitory computer-readable medium containing software code loadable into the internal memory of a computer having (a) a core part including logic;

• (b) a user interface to enable input and output to the user;
• and (c) a system interface to facilitate interaction of the core part with one or more software packages, applicable to Supply Chain Management in a multi-commodity, time-dependent setting wherein the number of constraints is typically smaller than the number of parameters, with one or more classes of uncertainty, with inputs being specified in one or more stages, having (i) one or more objective functions, each of the objective functions having one or more constraints, wherein the number of non-positivity constraints is typically smaller than the number of parameters, with one or more classes of uncertainty, (ii) input data represented in one or more ensembles over parameters with inputs being specified in one or more stages (iii) one or more assumptions about a future, and (v) quantitative estimates of information within the input data and one or more solutions being generated by:
• a. An interface configured to receive a hierarchical series of sets of constraints on parameters, with the parameters restricted to a constraint set forming an ensemble, and the hierarchy of constraints, represented as mathematical sets forming a hierarchy of ensemble, said hierarchy being based on subset, intersection or disjoint relationships by creating an ensemble with a specific optimal structure for supply chain management including nominal values for variables such as demand and input obtained in real-time by the computer;
• b. A constraint manager which utilizes optimization techniques to identify minimum and maximum bounds on each objective function, said bounds depending on the constraints comprising each ensemble of parameters and being computed for each of the assumptions about the future;
• c. An interface configured to display a volume metric of candidate ensembles and relating the volume to one or more information theoretic measures;
• d. logic responsive to information theoretic measures that adjusts the ensembles by changing the specificity of bounds on the objective function;
• e. logic arranged to quantify the amount of uncertainty in one or more constrained sets of parameters by identifying volume of constraint regions as information content;
• f. logic arranged to generate a tree of solutions by modifying constraints wherein solutions obtained under supersets of constraints are children of solutions obtained under their subsets;
• g. logic arranged to compute the degree of variation of outputs keeping the total input volume and uncertainty the same;
• h. logic arranged to transform the supply chain management structure to simulate perturbations on the variables within the optimal structure;
• and
• i. logic arranged to maximize or minimize a particular variable within the supply chain structure using linear programming.

A computer program product embodied in a non-transitory computer-readable medium containing software code loadable into the internal memory of a computer having (a) a core part including logic;

• (b) a user interface to enable input and output to the user;
• and (c) a system interface to facilitate interaction of the core part with one or more software packages, applicable to Analytical Hierarchical Processes, wherein the number of constraints is typically smaller than the number of parameters, with one or more classes of uncertainty, with inputs being specified in one or more stages, having (i) one or more objective functions, each of the objective functions having one or more constraints, wherein the number of constraints is typically smaller than the number of parameters, with one or more classes of uncertainty, (ii) input data represented in one or more ensembles with inputs being specified in one or more stages (iii) parameters based on the ensembles, (iv) one or more assumptions about a future, and (v) quantitative estimates of information within the input data to obtain a one or more rankings being generated by:
• a. An interface configured to receive a hierarchical series of sets of constraints on parameters, with the parameters restricted to a constraint set forming an ensemble, and the hierarchy of constraints, represented as mathematical sets forming a hierarchy of ensemble, said hierarchy being based on subset, intersection or disjoint relationships by creating an ensemble with a specific optimal structure for supply chain management including nominal values for variables such as demand and input obtained in real-time by the computer;
• b. A constraint manager which utilizes optimization techniques to identify minimum and maximum bounds on each objective function, said bounds depending on the constraints comprising each ensemble of parameters and being computed for each of the assumptions about the future;
• c. An interface configured to display a volume metric of candidate ensembles and relating the volume to one or more information theoretic measures;
• d. logic responsive to information theoretic measures that adjusts the ensembles by changing the specificity of bounds on the objective function;
• e. logic arranged to quantify the amount of uncertainty in one or more constrained sets of parameters by identifying volume of constraint regions as information content;
• f. logic arranged to generate a tree of solutions by modifying constraints wherein solutions obtained under supersets of constraints are children of solutions obtained under their subsets;
• g. logic arranged to compute the degree of variation of outputs keeping the total input volume and uncertainty the same;
• h. an interface arranged to receive user-specified relative ratios and alternatives for each criteria;
• and
• i. logic arranged to generate a random matrix to satisfy a constraint set specified by the user.

A computer program product embodied in a non-transitory computer-readable medium containing software code loadable into the internal memory of a computer having (a) a core part including logic;

• (b) a user interface to enable input and output to a user;
• and (c) a system interface to facilitate interaction of the core part with one or more software packages, applicable to Auctions, wherein the number of constraints is typically smaller than the number of parameters, with one or more classes of uncertainty, with inputs being specified in one or more stages, having (i) one or more objective functions, each of the objective functions having one or more constraints, wherein the number of constraints is typically smaller than the number of parameters, with one or more classes of uncertainty, (ii) input data represented in one or more ensembles with inputs being specified in one or more stages (iii) parameters based on the ensembles, (iv) one or more assumptions about a future, and (v) quantitative estimates of information within the input data, wherein the auctions rely on a valuation benefit function, which measures the benefit being generated by:
• a. An interface configured to receive a hierarchical series of sets of constraints on parameters, with the parameters restricted to a constraint set forming an ensemble, and the hierarchy of constraints, represented as mathematical sets forming a hierarchy of ensemble, said hierarchy being based on subset, intersection or disjoint relationships by creating an ensemble with a specific optimal structure for supply chain management including nominal values for variables such as demand and input obtained in real-time by the computer;
• b. A constraint manager which utilizes optimization techniques to identify minimum and maximum bounds on each objective function, said bounds depending on the constraints comprising each ensemble of parameters and being computed for each of the assumptions about the future;
• c. An interface configured to display a volume metric of candidate ensembles and relating the volume to one or more information theoretic measures;
• d. logic responsive to information theoretic measures to analyze and iteratively refine the ensembles by changing the specificity of bounds on the objective function;
• e. logic arranged to quantify the amount of uncertainty in one or more constrained sets of parameters by identifying volume of constraint regions as information content;
• f. logic arranged to generate a tree of solutions by modifying constraints wherein solutions obtained under supersets of constraints are children of solutions obtained under their subsets;
• g. logic arranged to compute the degree of variation of outputs keeping the total input volume and uncertainty the same;
• h. an interface arranged to receive user-specified relative ratios and alternatives for each criteria;
• and
• i. logic arranged to generate a random matrix to satisfy a constraint set specified by the user.