Image-based analytics of supply chain optimization problems
A system and method are disclosed for image analysis of supply chain planning problems modeled as a linear programming (LP) problems. Embodiments include receiving an LP matrix representing constraints and variables of at least a portion of a supply chain planning problem, generating a sorted variable index for the variables of the LP matrix, generating a sorted constraint index for the constraints of the LP matrix, identifying functions of the variables and the constraints, selecting one or more colors of pixels of a supply chain problem image based, at least in part, on the identified functions of the variables and constraints, selecting locations of the pixels, and displaying a visualization of the supply chain problem image.
1. A system of image analysis of supply chain planning problems modeled as linear programing (LP) problems, comprising:
a computer, comprising a processor and memory, the computer configured to:
receive an LP matrix representing constraints and variables of at least a portion of a supply chain planning problem;
generate a sorted variable index for the variables of the LP matrix;
generate a sorted constraint index for the constraints of the LP matrix;
identify functions of the variables and the constraints;
color pixels of a supply chain problem image based, at least in part, on logic that maps a link between each pixel and a corresponding individual variable or constraint to determine a color of each pixel;
determine locations of the pixels on the supply chain problem image based, at least in part, on a mapping between each pixel and a column matrix location of the sorted variable index and a row matrix location of the sorted constraint index; and
display a visualization of the supply chain problem image comprising the colored pixels located at the determined locations.
2. The system of claim 1 , wherein the LP matrix represents, at least in part, an LP formulation of a static structure and dynamic properties of a supply chain.
3. The system of claim 1 , wherein the computer is further configured to:
represent nodes of a supply chain as space-time sorted nodes, each of the space-time sorted nodes associated with a supply chain stream location and a time bucket, wherein the sorted variable index indexes each of the variables according to the supply chain stream location and the time bucket of the space-time sorted nodes, and the sorted constraint index indexes each of the constraints according to the supply chain stream location and the time bucket of the space-time sorted nodes.
4. The system of claim 3 , wherein the nodes of the supply chain comprise one or more of a material buffer, a capacity, a manufacturing process, and a distribution process.
5. The system of claim 1 , wherein the visualization comprises a structural image analysis of the LP matrix, the visualization comprising clusters and one or more contours, wherein at least two of the clusters represent supply chain components and at least one of the one or more contours represents a flow connecting the supply chain components.
6. The system of claim 1 , wherein the visualization comprises a supply chain structure and complexity visualization, wherein the color of the one or more pixels is based, at least in part, on a domain and a categorization of a supply chain element.
7. The system of claim 1 , wherein the LP matrix is a first LP matrix generated prior to solving a first objective and the computer is further configured to:
receive a second LP matrix generated prior to solving a second objective, wherein the visualization comprises a first supply chain element that is critical to the first objective that is visually distinguishable from a second supply chain element that is critical to the second objective.
8. A computer-implemented method of image analysis of supply chain planning problems modeled as linear programing (LP) problems, comprising:
receiving an LP matrix representing constraints and variables of at least a portion of a supply chain planning problem;
generating a sorted variable index for the variables of the LP matrix;
generating a sorted constraint index for the constraints of the LP matrix;
identifying functions of the variables and the constraints;
coloring pixels of a supply chain problem image based, at least in part, on logic that maps a link between each pixel and a corresponding individual variable or constraint to determine a color of each pixel;
determining locations of the pixels on the supply chain problem image based, at least in part, on a mapping between each pixel and a column matrix location of the sorted variable index and a row matrix location of the sorted constraint index; and
displaying a visualization of the supply chain problem image comprising the colored pixels located at the determined locations.
9. The computer-implemented method of claim 8 , wherein the LP matrix represents, at least in part, an LP formulation of a static structure and dynamic properties of a supply chain.
10. The computer-implemented method of claim 8 , further comprising:
representing nodes of a supply chain as space-time sorted nodes, each of the space-time sorted nodes associated with a supply chain stream location and a time bucket, wherein the sorted variable index indexes each of the variables according to the supply chain stream location and the time bucket of the space-time sorted nodes, and the sorted constraint index indexes each of the constraints according to the supply chain stream location and the time bucket of the space-time sorted nodes.
11. The computer-implemented method of claim 10 , wherein the nodes of the supply chain comprise one or more of a material buffer, a capacity, a manufacturing process, and a distribution process.
12. The computer-implemented method of claim 8 , wherein the visualization comprises a structural image analysis of the LP matrix, the visualization comprising clusters and one or more contours, wherein at least two of the clusters represent supply chain components and at least one of the one or more contours represents a flow connecting the supply chain components.
13. The computer-implemented method of claim 8 , wherein the visualization comprises a supply chain structure and complexity visualization, wherein the color of the one or more pixels is based, at least in part, on a domain and a categorization of a supply chain element.
14. The computer-implemented method of claim 8 , wherein the LP matrix is a first LP matrix generated prior to solving a first objective and the computer-implemented method further comprises:
receiving a second LP matrix generated prior to solving a second objective, wherein the visualization comprises a first supply chain element that is critical to the first objective that is visually distinguishable from a second supply chain element that is critical to the second objective.
15. A non-transitory computer-readable medium embodied with software providing image analysis of supply chain planning problems modeled as linear programing (LP) problems, the software when executed:
receives an LP matrix representing constraints and variables of at least a portion of a supply chain planning problem;
generates a sorted variable index for the variables of the LP matrix;
generate a sorted constraint index for the constraints of the LP matrix;
identifies functions of the variables and the constraints;
colors pixels of a supply chain problem image based, at least in part, on logic that maps a link between each pixel and a corresponding individual variable or constraint to determine a color of each pixel;
determines locations of the pixels on the supply chain problem image based, at least in part, on a mapping between each pixel and a column matrix location of the sorted variable index and a row matrix location of the sorted constraint index; and
displays a visualization of the supply chain problem image comprising the colored pixels located at the determined locations.
16. The non-transitory computer-readable medium of claim 15 , wherein the LP matrix represents, at least in part, an LP formulation of a static structure and dynamic properties of a supply chain.
17. The non-transitory computer-readable medium of claim 15 , wherein the software when executed further:
represents nodes of a supply chain as space-time sorted nodes, each of the space-time sorted nodes associated with a supply chain stream location and a time bucket, wherein the sorted variable index indexes each of the variables according to the supply chain stream location and the time bucket of the space-time sorted nodes, and the sorted constraint index indexes each of the constraints according to the supply chain stream location and the time bucket of the space-time sorted nodes.
18. The non-transitory computer-readable medium of claim 15 , wherein the visualization comprises a structural image analysis of the LP matrix, the visualization comprising clusters and one or more contours, wherein at least two of the clusters represent supply chain components and at least one of the one or more contours represents a flow connecting the supply chain components.
19. The non-transitory computer-readable medium of claim 15 , wherein the visualization comprises a supply chain structure and complexity visualization, wherein the color of the one or more pixels is based, at least in part, on a domain and a categorization of a supply chain element.
20. The non-transitory computer-readable medium of claim 15 , wherein the LP matrix is a first LP matrix generated prior to solving a first objective and the software when executed further:
receives a second LP matrix generated prior to solving a second objective, wherein the visualization comprises a first supply chain element that is critical to the first objective that is visually distinguishable from a second supply chain element that is critical to the second objective.