Generation and use of a data structure in operational planning of configured products
The use of a DAG data structure to represent valid product configurations addresses the complexity of operational planning by encoding valid variants and enabling efficient column generation, facilitating effective long-term sales option-based planning and cost calculations.
Patent Information
- Application Number
- PCT/IB2024/054552
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-11
- Filing Date
- 2024-05-10
- Publication Date
- 2024-11-14
AI Technical Summary
The complexity of product configuration in operational planning, where numerous characteristics and variables lead to an insurmountable number of configurations, making it difficult for computers to efficiently determine valid configurations and optimize planning.
The generation and use of a directed acyclic graph (DAG) data structure, which encodes valid variants of configured products based on a ruleset, allowing for the representation of all valid combinations and enabling column generation for linear programming, even when the number of variants is large, thereby solving the planning challenges.
This approach allows for the efficient generation and use of valid product configurations, enabling effective long-term sales option-based planning and cost calculations, and determining more profitable variants, by representing valid combinations in a structured format that can be processed without generating the entire DAG.
Smart Images

Figure IB2024054552_14112024_PF_FP_ABST
Abstract
Description
GENERATION AND USE OF A DATA STRUCTURE IN OPERATIONAL PLANNING OF CONFIGURED PRODUCTSTECHNICAL FIELD
[0001] The present disclosure relates to systems for generating a data structure, such as a directed acyclic graph (DAG), in operational planning of configured products.BACKGROUND
[0002] The butterfly effect, often associated with chaos theory, can be understood as a small change in an initial or early state of a deterministic nonlinear system causing large differences in a later state. Often, the butterfly effect is associated with the work of Edward Norton Lorenz, a renowned meteorologist and mathematician, with his teachings on the formation of tornados. He explained that the time of formation and path taken can be influenced by small perturbations such as a distant butterfly flapping its wings weeks earlier. Lorenz suggested that a weather model fails to reproduce the results of runs with unrounded initial condition data in that a minute change in initial conditions can produce a substantially different result.
[0003] Beyond weather modeling, the butterfly effect concept has since been used as a broad term for any situation where a small change may be the cause of larger consequences— which leads us to a brief discussion of its effect on the planning of configured products. The butterfly effect can have a profound influence on a significant problem in operational planning of configured products. Often with a product configuration, there are tens if not hundreds and sometimes even thousands of characteristics in one model of a product’s variations. And, when multiple variables occur for each characteristic, it is to be understood that the number of combinations or configurations becomes insurmountable for even a contemporary computer to consider in a practical amount of time. E.g., see FIG. 1, which illustrates how increasing variables or characteristics in a situation for product configuration by merely three dimensions leads to a drastic non-linear increase in variables or characteristics.
[0004] Also, there are difficulties in understanding valid configurations in a model with multiple characteristics and multiple variables per characteristic. FIG. 2 shows that the amount of valid configurations in a model can be significant and thus selecting such configurations does not immediately solve the problem of insurmountable numbers in the operational planning of configured products. For example, given up to 1200 characteristics in a configured product, with an average of 6 variables per characteristic, the number of configurations is staggering (e.g., approximately 10A583). And, understanding and choosing valid configurations from such a situation can also be a significant and technical problem given that computers are not capable of solving problems with such a great number of alternative situations. In other words, the great complexity in resolving operational planning of configured products objectively and optimally is a significant technical problem to be resolved with one or more technical solutions.
[0005] FIG. 3 shows an imprint or data cloud of feasible combinations in three dimensions and from a specified direction and using a filter to show the extremely large amount of combinations that occur in the operational planning of a configured product.SUMMARY
[0006] Described herein are techniques for generation and use of a data structure, such as a DAG, in operational planning of configured products. The techniques disclosed herein provide specific technical solutions to at least overcome the technical problems mentioned in the background section or other parts of the application as well as other technical problems not described herein but recognized by those skilled in the art. In some embodiments, the techniques include technologies for generation and use of a data structure for long-term sales option-based planning of configured products (such as configured agricultural products or farming machines). The long-term sales option-based planning can include sales option-based planning and the configured products can include complex models of products. In some embodiments, the technologies include a method that includes generating a data structure, such as a DAG, based on a ruleset for variants of a configured product (such as a configured agricultural product or farming machine), wherein the data structure includes valid variants of the configured product. In some embodiments using a DAG as the data structure, traversals of the DAG represent valid combinations according to the ruleset (such as allvalid combinations), and each traversal of the DAG from a root node of the DAG to a terminal node of the DAG represents a single valid variant of the configured product. In some embodiments, a linear program (LP) is applied to the operational planning of configured products. For example, the LP can be used for long-term sales option-based planning, sales option cost calculations, unique product recommendations, and determining more profitable variants of a product. The data structure (such as a DAG) represents the valid combinations even when the number of variants is very large and it is a basis of column generation for the LP using features of the data structure. In some examples, a DAG represents the valid combinations even when the number of variants is very large and it is a basis of column generation for the LP using the traversals (making it possible to solve the LP without generating the DAG completely).
[0007] With respect to some embodiments, disclosed herein are computerized methods for generating and using a data structure (such as a DAG) in operational planning of configured products (such as configured agricultural or farming products) as well as a non-transitory computer-readable storage medium for carrying out technical operations of the computerized methods. The non- transitory computer-readable storage medium has tangibly stored thereon, or tangibly encoded thereon, computer-readable instructions that when executed by one or more devices (e.g., one or more personal computers or servers) cause at least one processor to perform a method for improved systems and methods for generating and using a data structure in operational planning of configured products.
[0008] For example, some embodiments include a method for generating and using a data structure, e.g., a DAG, in operational planning of configured products, e.g., configured agricultural or farming products (e.g., see method 300 shown in FIG. 6). In some examples, the method includes formulating, by a computing system (e.g., see computing system 200 shown in FIG. 5), an LP to solve a technical problem about variants of a configured agricultural product permitted by a given ruleset (e.g., see step 302 of method 300 shown in FIG. 6). The LP is formulated with a linear objective function on variables that represent characteristic values of the product. The method also includes generating, by the computing system, a data structure (such as a DAG) to determine at least some of the variants of the product (e.g., see step 304 of method 300). The data structure (such as the DAG) encodes valid variants according to the given ruleset. In some embodiments using the DAG, each traversal of the DAG from the root node of the DAG to a terminal node of the DAGrepresents a single valid variant of the configured agricultural product. Also, a level of the DAG corresponds to a specific characteristic and an edge in a traversal of the DAG corresponds to a choice of a value for an associated characteristic. The method also includes generating, by the computing system, columns for the LP according to the data structure, such as the DAG (e.g., see step 306).
[0009] In some embodiments, the method includes displaying the data structure (e.g., the DAG) in a graphical user interface (GUI) (e.g., see step 308). In some examples, the method includes storing in memory, by the computing system, the data structure (e.g., the DAG), wherein the data structure is an uncompressed data structure (e.g., see step 310). Also, the method can include compressing, by the computing system, the stored data structure (e.g., see step 312) as well as displaying the uncompressed data structure (e.g., see uncompressed DAG 900 shown in FIG. 9) along with the compressed data structure (e.g., see compressed DAG 1000 shown in FIG. 10) in the GUI (e.g., see step 314a).
[0010] In some embodiments, solving the UP requires the generation of the columns. In some embodiments, the data structure (e.g., the DAG) is generated with a backtracking algorithm with use of a ruleset satisfiability oracle (e.g., see step 402 shown in FIG. 7A). In some embodiments, the data structure (e.g., the DAG) is generated from a Boolean logic formula derived from the given ruleset, such as with use of an SAT solver (e.g., see steps 502 and 504 depicted in FIG. 7B). In some embodiments, the generation of the columns is based on Dantzig-Wolfe decomposition. In some cases, the generation of the columns is based on traversals of the data structure (e.g., the DAG). And, in some examples, the method includes, in solving the UP, delaying column generation to improve tractability of the UP.
[0011] In some embodiments, the method includes generating a characteristic dependency graph based on the given ruleset (e.g., see step 602 shown in FIG. 7C). Also, an order of levels of the data structure (e.g., the DAG) is at least partially based on the characteristic dependency graph (e.g., see step 604). And, the method further includes compressing the data structure (e.g., the DAG) based on the characteristic dependency graph (e.g., see step 606).
[0012] In some embodiments, all valid combinations are known and each of the valid combinations has an associated variable in the UP (e.g., see step 808 shown in FIG. 8), and when all of the valid combinations of a given configuration are represented in the DAG (e.g., see step 808),the computing system (e.g., see computing system 200) uses the data structure (e.g., the DAG) in the generation of the columns to solve the LP and to resolve having a very large number of combinations which is associated with the technical problem (e.g., see step 810). In some embodiments, the method includes determining, by the computing system, a suggestion for changing input parameters for the LP in a least intrusive way according to a target function, when the LP has no feasible solution (e.g., see step 812). And, the method can include using, by the computing system, historical data to determine a solution that minimizes total deviation from past component consumption, when the LP has multiple feasible solutions (e.g., see step 814).
[0013] In some embodiments, the method includes repeating the generating of the data structure (e.g., the DAG) and the formulating of the LP, for the given ruleset, multiple times to further enhance a final DAG outputted by the method. The method can also include, in the repeating of the generating of the data structure: storing each generated instance of the data structure so that a multiple of the data structures are retrievable (e.g., multiple retrieval DAGs); determining results of the given ruleset initially, before the generation of the multiple data structures (e.g. , multiple DAGs); and determining a sequence of nodes of the multiple data structures according to repetition in attributes of the results of the given ruleset initially before the generation of the multiple data structures.
[0014] In some embodiments, the configured product is or includes a configured farming machine, configured agricultural software, a configured hardware system for a farming machine, a configured software system for a farming machine, a configured system including farming machine hardware and software that supports or controls the hardware, or any combination thereof. In some embodiments, the method is for long-term sales option-based planning. The long-term sales optionbased planning can include sales option-based planning of configured products (such as configured agricultural or farming products), and the configured products can include complex models of products (such as configured agricultural or farming products).
[0015] In some embodiments, the method includes using a module of an enterprise resource planning (ERP) system to generate the ruleset (e.g., see step 702 shown in FIG. 7D). The method can also include using the module of the ERP system to generate the valid variants based on the generated ruleset (e.g., see step 704). And, the method can include using the generated valid variantsto generate the data structure, e.g., the DAG (e.g., see step 706). Also, the method can be integrated and implemented through the ERP system.
[0016] And, in some embodiments, the method can include compressing the data structure (e.g., the DAG) vertically, horizontally, or any combination thereof. And, the method can include selecting a sub-graph of the data structure (e.g., the DAG) and compressing the selected sub-graph instead of the entirety of the data structure.
[0017] With respect to some embodiments, a system is provided that includes at least one computing device configured to provide improved ways for generating and using a data structure (such as a DAG) in operational planning of configured products (such as configured agricultural or farming products). And, with respect to some embodiments, a method, such as one of the aforesaid methods, is provided to be performed by at least one computing device. In some example embodiments, computer program code can be executed by at least one processor of one or more computing devices to implement functionality in accordance with at least some embodiments described herein; and the computer program code being at least a part of or stored in a non-transitory computer-readable medium.
[0018] For example, some embodiments are or include a system for generating and using a data structure, e.g., a DAG, in operational planning of configured products, e.g., configured agricultural or farming products (e.g., see computing system 200). The system includes a processing device (e.g., see processing device 202) and memory (e.g., see memory 204) in communication with the processing device and storing instructions (e.g., see instructions 214) that, when executed by the processing device, cause the processing device to provide functionality described herein for overcoming the technical problems due to the butterfly effect and other related problems in operational planning of configured products. The instructions when executed can cause the system to formulate an LP to solve a technical problem about variants of a configured product (such as an agricultural or farming product) permitted by a given ruleset (e.g., see instructions 220). The LP is formulated with a linear objective function on variables that represent characteristic values of the product. The instructions when executed can also cause the system to generate a data structure to determine at least some of the variants of the product (e.g., see instructions 222). The data structure encodes valid variants according to the given ruleset, and each iterative use of the data structure represents a single valid variant of the configured product. The instructions when executed can alsocause the system to generate columns for the LP according to the data structure (e.g., see instructions 224).
[0019] The systems and methods described herein overcome some technical problems in operational planning of configured products, in general, as well as some technical problems in operational planning of configured agricultural or farming products. Also, the techniques disclosed herein provide specific technical solutions to at least overcome the technical problems mentioned in the background section or other parts of the application as well as other technical problems not described herein but recognized by those skilled in the art.
[0020] These and other important aspects of the invention are described more fully in the detailed description below. The invention is not limited to the particular methods and systems described herein. Other embodiments can be used and changes to the described embodiments can be made without departing from the scope of the claims that follow the detailed description.
[0021] Within the scope of this application, it should be understood that the various aspects, embodiments, examples, and alternatives set out herein, and individual features thereof may be taken independently or in any possible and compatible combination. Where features are described with reference to a single aspect or embodiment, it should be understood that such features are applicable to all aspects and embodiments unless otherwise stated or where such features are incompatible.BRIEF DESCRIPTION OF THE DRAWINGS
[0022] The present disclosure will be understood more fully from the detailed description given below and from the accompanying drawings of various embodiments of the disclosure. Embodiments of the invention will now be described, by way of example, with reference to the accompanying drawings, in which:
[0023] FIGS. 1 to 3 illustrate example technical problems to be solved in the operational planning of configured products;
[0024] FIG. 4 illustrates an example technical solution to the example technical problems shown in FIGS. 1 to 3, in accordance with some embodiments of the present disclosure;
[0025] FIG. 5 illustrates is a block diagram of example aspects of a computing system, in accordance with some embodiments of the present disclosure;
[0026] FIGS. 6, 7A, 7B, 7C, 7D, and 8 illustrate methods in accordance with some embodiments of the present disclosure;
[0027] FIG. 9 illustrates an image of an uncompressed DAG displayed in a part of a GUI, in accordance with some embodiments of the present disclosure;
[0028] FIG. 10 illustrates an image of a compressed DAG displayed in a part of a GUI, which is at least compressed vertically, in accordance with some embodiments of the present disclosure; and
[0029] FIG. 11 illustrates images of an uncompressed DAG and a horizontally compressed DAG displayed in respective parts of a GUI, in accordance with some embodiments of the present disclosure.DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS
[0030] Details of example embodiments of the invention are described in the following detailed description with reference to the drawings. Although the detailed description provides reference to example embodiments, it is to be understood that the invention disclosed herein is not limited to such example embodiments. But to the contrary, the invention disclosed herein includes numerous alternatives, modifications, and equivalents as will become apparent from consideration of the following detailed description and other parts of this disclosure.
[0031] In general, described herein are technologies for generation and use of a data structure in operational planning of configured products. In some embodiments, the configured products include configured farming or agricultural products. In some embodiments, the technologies include generation and use of the data structure for long-term sales option-based planning of configured products (e.g., configured farming machines). The long-term sales option-based planning can include sales option-based planning and the configured products can include complex models of products. In some embodiments, the technologies include a method that includes generating the data structure based on a ruleset for variants of a configured product, wherein the data structure includesvalid variants of the product. The traversals of the data structure (such as when the structure is a DAG) can represent the valid combinations according to the given ruleset (such as all valid combinations). In some embodiments, a linear program (LP) is applied to the operational planning. For example, the LP can be used for long-term sales option-based planning, sales option cost calculations, unique product recommendations, and determining more profitable variants of a product. The data structure represents the valid combinations even when the number of variants is very large and it is a basis of column generation for the LP using the traversals (making it possible to solve the LP without generating the DAG completely).
[0032] Some embodiments include technologies for generation and use of a DAG in operational planning of configured products. In some embodiments, the technologies include generation and use of a DAG for long-term sales option-based planning of configured agricultural products (e.g., configured farming machines). The long-term sales option-based planning can include sales optionbased planning and the agricultural products can include complex models of agricultural products. In some embodiments, the technologies include a method that includes generating a DAG based on a ruleset for variants of a configured agricultural product, wherein the DAG includes valid variants of the product. The traversals of the DAG represent the valid combinations according to the given ruleset (such as all valid combinations), and each traversal of the DAG from a root node of the DAG to a terminal node of the DAG represents a single valid variant of the product. In some embodiments, a linear program (LP) is applied to the agricultural operations. For example, the LP can be used for long-term sales option-based planning, sales option cost calculations, unique product recommendations, and determining more profitable variants of a product. The DAG represents the valid combinations even when the number of variants is very large and it is a basis of column generation for the LP using the traversals (making it possible to solve the LP without generating the DAG completely).
[0033] FIG. 4 illustrates an example technical solution 100 to the example technical problems shown through FIGS. 1 to 3. The technical solution can include or be a part of the techniques and technologies described herein. As suggested by FIGS. 1 to 3, beyond weather modeling, the butterfly effect can have a profound influence on a significant problem in operational planning of configured products. Often with a product configuration, there are tens if not hundreds and sometimes even thousands of characteristics in one model of a product’s variations. And, when multiple variablesoccur for each characteristic, it is to be understood that the number of combinations or configurations becomes insurmountable for even a contemporary computer to consider. FIG. 1 shows how increasing a situation for product configuration by merely three dimensions leads to a drastic non-linear increase in variables or characteristics. Also, there are difficulties in understanding valid configurations in a model with multiple characteristics and multiple variables per characteristic. FIG. 2 shows that the amount of valid configurations in a model can be significant and thus selecting such configurations does not immediately solve the problem of insurmountable numbers in the operational planning of configured products. For example, as shown in FIGS. 1 and 2, given up to 1200 characteristics in a configured product, with an average of 6 variables per characteristic, the number of configurations is staggering (e.g., approximately 10A583). FIG. 3 shows an imprint or data cloud of feasible combinations in three dimensions and from a specified direction and use of a filter to show the extremely large amount of combinations that occur in the operational planning of a configured product. Understanding and choosing valid configurations from such a situation can also be a significant and technical problem given that computers are not capable of solving problems with such a great number of alternative situations. In other words, the great complexity in resolving operational planning of configured products objectively and optimally is a significant technical problem to be resolved with one or more technical solutions (such as the technical solutions provided in FIGS. 4 to 8).
[0034] The technical solution 100, shown in FIG. 4, can include or be a part of the techniques and technologies described herein (such as any one of the methods 300, 600, 700, and 800 shown in FIGS. 6, 7A, 7B, 7C, 7D, and 8 respectively) and can provide specific technical solutions to at least overcome the technical problems mentioned in the background section or other parts of the application as well as other technical problems not described herein but recognized by those skilled in the art. For example, a universal variant configuration enabler 102 (UVCE 102) through data analysis can provide a visualization 104 of one or more solutions for the technical problem of the butterfly effect causing tens if not hundreds and sometimes even thousands of characteristics in one model of a product’s variations. And, when multiple variables occur for each characteristic, it is to be understood that the number of combinations or configurations becomes insurmountable for even a contemporary computer to consider due to the butterfly effect. However, the UVCE 102 can provide the visualization 104 to at least partially overcome such problems. The visualization 104can include one or more images of DAGs, Hamming graphs, and other types of data structures. The visualization 104 as well as the UVCE 102 can be prerequisites or a basis for determinations and derivations of data used with the applications of the UVCE 106. The applications of the UVCE 106 can include utilizing the techniques for generation and use of a data structure, such as a DAG, in operational planning of configured products as well as technologies for generation and use of a data structure for long-term sales option-based planning of configured products (such as configured agricultural products or farming machines). The long-term sales option-based planning can include sales option-based planning and the configured products can include complex models of products.
[0035] FIG. 5 illustrates is a block diagram of example aspects of a computing system 200 that can implement the technical solution 100 or each main part of the solution, such as the UVCE 102, the visualization 104, and the applications of the UVCE 106. Also, FIG. 2 illustrates parts of the computing system 200 within which a set of instructions are executed for causing a machine (such as a computer processor or processing device 202) to perform any one or more of the methodologies discussed herein performed by a computing system (e.g., see the method steps of the methods 300, 600, 700, and 800 shown in FIGS. 6, 7A, 7B, 7C, 7D, and 8 respectively). In some embodiments, the computing system 200 operates with additional computing systems to provide increased computing capacity in which multiple computing systems operate together to perform any one or more of the methodologies discussed herein that are performed by a computing system.
[0036] In some embodiments, the computing system 200 corresponds to a host system that includes, is coupled to, or utilizes memory or is used to perform the operations performed by any one of the computing systems described herein. In some embodiments, the machine is connected (e.g., networked) to other machines in a LAN, an intranet, an extranet, or the Internet. In some embodiments, the machine operates in the capacity of a server in a client-server network environment, as a peer machine in a peer-to-peer (or distributed) network environment, or as a server in a cloud computing infrastructure or environment. In some embodiments, the machine is a personal computer (PC), a tablet PC, a cellular telephone, a web appliance, a server, or any machine capable of executing a set of instructions (sequential or otherwise) that specify actions to be taken by that machine. Further, while a single machine is illustrated, the term “machine” shall also be taken to include any collection of machines that individually or jointly execute a set (or multiple sets) ofinstructions to perform any one or more of the methodologies discussed herein performed by computing systems.
[0037] The computing system 200 includes a processing device 202, a main memory 204 (e.g., read-only memory (ROM), flash memory, dynamic random-access memory (DRAM), etc.), a static memory 206 (e.g., flash memory, static random-access memory (SRAM), etc.), and a data storage system 210, which communicate with each other via a bus 218. The processing device 202 represents one or more general-purpose processing devices such as a microprocessor, a central processing unit, or the like. More particularly, the processing device can include a microprocessor or a processor implementing other instruction sets, or processors implementing a combination of instruction sets. Or, the processing device 202 is one or more special-purpose processing devices such as an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), a network processor, or the like. The processing device 202 is configured to execute instructions 214 for performing the operations discussed herein performed by a computing system. In some embodiments, the computing system 200 includes a network interface device 208 to communicate over a communications network. Such a communications network can include one or more local area networks (LAN(s)) and / or one or more wide area networks (WAN(s)). In some embodiments, the communications network includes the Internet and / or any other type of interconnected communications network. The communications network can also include a single computer network or a telecommunications network. More specifically, in some embodiments, the communications network includes a local area network (LAN) such as a private computer network that connects computers in small physical areas, a wide area network (WAN) to connect computers located in different geographical locations, and / or a middle area network (MAN) to connect computers in a geographic area larger than that covered by a large LAN but smaller than the area covered by a WAN.
[0038] The data storage system 210 includes a machine-readable storage medium 212 (also known as a computer-readable medium) on which is stored one or more sets of instructions 214 or software embodying any one or more of the methodologies or functions described herein performed by a computing system. The instructions 214 also reside, completely or at least partially, within the main memory 204 or within the processing device 202 during execution thereof by the computingsystem 200, the main memory 204 and the processing device 202 also constituting machine-readable storage media.
[0039] In some embodiments, as illustrated in FIG. 5, the instructions 214 include specific instructions to implement functionality described herein related to the methods described herein and that can correspond to any one of the computing devices, data processors, UI devices, and I / O devices described herein related to a computing system. For example, the instructions 214 include linear program formulation instructions 220. The instructions 220 are executable by the processing device 202 to cause the computing system 200 to formulate a linear program (LP) to solve a technical problem about variants of a configured product (such as an agricultural or farming configured product) permitted by a given ruleset. The LP is formulated with a linear objective function on variables that represent characteristic values of the product. In some embodiments, the data structure is generated with a backtracking algorithm with use of a ruleset satisfiability oracle. In some embodiments, the data structure is generated from a Boolean logic formula derived from the given ruleset, such as with use of an SAT solver. In some examples, the generation of the columns is based on Dantzig-Wolfe decomposition. Also, in some examples, the generation of the columns is based on traversals of the data structure.
[0040] In some embodiments, the configured product is a configured agricultural or farming product. And, in some examples, the product is or includes a configured farming machine, a configured agricultural software, a configured hardware system for a farming machine, a configured software system for a farming machine, a configured system including farming machine hardware and software that supports or controls the hardware, or any combination thereof.
[0041] Also, the instructions 214 include variant encoding instructions 222 that are executable by the processing device 202 to cause the computing system 200 to generate a data structure to determine at least some of the variants of the product. The data structure encodes valid variants according to the given ruleset, and each iterative use of the data structure represents a single valid variant of the configured product. In some examples, the data structure includes a graph. And, in some instances, the graph includes a DAG or another type of graph.
[0042] The instructions 214 also include column generation instructions 224 that are executable by the processing device 202 to cause the computing system 200 to generate columns for the LPaccording to the data structure. In some embodiments, solving the LP requires the generation of the columns. And, in solving the LP, the computing system, via the instructions 214 can delay the column generation to improve tractability of the LP.
[0043] The instructions 214 further include visualization instructions 226 that are executable by the processing device 202 to cause the computing system 200 to display the data structure in a GUI. Also, the instructions 214 include compression instructions 228 that are executable by the processing device 202 to cause the computing system 200 to compress the data structure before or after it has been stored the by the computing system 200. And, the visualization instructions 226 are further executable by the processing device 202 to cause the computing system 200 to display the uncompressed data structure along with the compressed data structure in the GUI, such as through the UI 216. E.g., see the displaying of an uncompressed DAG 900 shown in FIG. 9 and a compressed DAG 1000 shown in FIG. 10.
[0044] In some embodiments, the variant encoding instructions 222 are executable by the processing device 202 to cause the computing system 200 to generate a characteristic dependency graph based on the given ruleset. In some of such cases, an order of levels of the data structure is at least partially based on the characteristic dependency graph. And, the compression instructions 228 can be further executable by the processing device 202 to cause the computing system 200 to compress the data structure based on the characteristic dependency graph.
[0045] In some embodiments, the data structure includes a DAG and the compression instructions 228 are executable by the processing device 202 to compress the DAG vertically, horizontally, or any combination thereof. And, these instructions or another part of the instructions 214 can be executed such the computing system selects a sub-graph of the DAG and compresses the selected sub-graph instead of the entirety of the DAG.
[0046] The instructions 214 also include product solutions instructions 230 that are executable by the processing device 202 to cause the computing system 200 to use the data structure in the generation of the columns to solve the LP and to resolve having a very large number of combinations which is associated with the technical problem. The use of the data structure in such a way occurs when all of the valid combinations of a given configuration are represented in the data structure and all valid combinations are known and each of the valid combinations has an associated variable inthe LP. The product solutions instructions 230 are also executable by the processing device 202 to cause the computing system 200 to determine a suggestion for changing input parameters for the LP in a least intrusive way according to a target function, when the LP has no feasible solution. And, the product solutions instructions 230 are also executable by the processing device 202 to cause the computing system 200 to use historical data to determine a solution that minimizes total deviation from past component consumption when the LP has multiple feasible solutions.
[0047] In some embodiments, the instructions 214 are executable by the processing device 202 to cause the computing system 200 to repeat the generation of the data structure and the formulation of the LP, for the given ruleset, multiple times to further enhance a final instance of the data structure outputted by the computing system 200. In the repeating of the generation of the data structure, the computing system 200 stores each generated instance of the data structure so that multiple data structures are retrievable and determine results of the given ruleset initially, before the generation of the multiple instances of the data structure via the variant encoding instructions 222. The computing system 200 is also configured, in accordance with the instructions 214, to determine a sequence of nodes or other types of features of the multiple instances of the data structure according to repetition in attributes of the results of the given ruleset initially before the generation of the multiple instances of the data structures.
[0048] In some embodiments, the computing system 200 is configured to, according to the instructions 214, apply the functionality described herein to long-term sales option-based planning. In some examples, the long-term sales option-based planning includes sales option-based planning of configured products. And, in some cases, the configured products include complex models of products.
[0049] Furthermore, the instructions 214 include ERP system integration instructions 232 that are executable by the processing device 202 to cause the computing system 200 to use a module of an enterprise resource planning (ERP) system to generate the ruleset. Also, the ERP system integration instructions 232 are executable by the processing device 202 to cause the computing system 200 to use the module of the ERP system to generate the valid variants based on the generated ruleset. And, the ERP system integration instructions 232 are executable by the processing device 202 to cause the computing system 200 to use the generated valid variants to generate the datastructure. In some embodiments, the instructions 214 are integrated and implemented through the ERP system.
[0050] While the machine-readable storage medium 212 is shown in an example embodiment to be a single medium, the term “machine-readable storage medium” should be taken to include a single medium or multiple media that store the one or more sets of instructions. The term “machine- readable storage medium” shall also be taken to include any medium that is capable of storing or encoding a set of instructions for execution by the machine and that causes the machine to perform any one or more of the methodologies of the present disclosure performed by a computing system. The term “machine-readable storage medium” shall accordingly be taken to include solid-state memories, optical media, or magnetic media.
[0051] Also, as shown, the computing system 200 includes user interface 216 that includes a display, in some embodiments, and, for example, implements functionality corresponding to any one of the UI devices disclosed herein. A UI, such as UI 216, or a UI device described herein includes any space or equipment where interactions between humans and machines occur. A UI described herein allows operation and control of the machine from a human user, while the machine simultaneously provides feedback information to the user. Examples of a user interface, or UI device include the interactive aspects of computer operating systems (such as GUIs), machinery operator controls, and process controls. A UI described herein includes one or more layers, including a human-machine interface (HMI) that interfaces machines with physical input hardware and output hardware.
[0052] In some systems of the technologies disclosed herein, any steps of embodiments of the methods described herein are implementable by executing instructions corresponding to the steps, which are stored in memory (e.g., see instructions 214, 220, 222, 224, 226, 228 and 230 shown in FIG. 5).
[0053]
[0054] FIGS. 6 to 8 illustrate methods in accordance with some embodiments of the present disclosure. Methods 300, 600, 700, and 800 of FIGS. 6 to 8, respectively, are performed by any one of the computing systems described herein (e.g., see computing system 200 depicted in FIG. 5). Themethods 600 and 700, shown in FIGS. 7C and 7D respectively, include the method step 304 of method 300 as well as the remainder of method 300 (which is not depicted).
[0055] As shown in FIG. 300, method 300 begins with step 302, which includes formulating, by a computing system, an LP to solve a technical problem about variants of a configured product (such as an agricultural or farming configured product) permitted by a given ruleset. The LP is formulated with a linear objective function on variables that represent characteristic values of the product. At step 304, the method 300 continues with the computing system generating a data structure to determine at least some of the variants of the product. The data structure encodes valid variants according to the given ruleset, and each iterative use of the data structure represents a single valid variant of the configured agricultural product. At step 306, the method 300 continues with generating, by the computing system, columns for the LP according to the data structure. In some examples, the computing system solves the LP and requires the generation of the columns to do the solving of the LP. In some examples, the generation of the columns is based on Dantzig-Wolfe decomposition. In some cases, the generation of the columns is based on traversals of the data structure. And, in some embodiments, the method 300 includes, in solving the LP, the computing system delaying column generation to improve tractability of the LP (not depicted).
[0056] At step 308, the method 300 continues with the computing system (such via its UI) displaying the data structure in a GUI. At step 310, the method continues with storing in memory, by the computing system, the data structure, wherein the data structure is an uncompressed data structure. At step 312, the method continues with compressing, by the computing system, the stored data structure. At step 314a, the method continues with displaying the uncompressed data structure along with the compressed data structure in the GUI. Step 314 shows that the computing system can display the compressed data structure with or without the uncompressed data structure. E.g., see the displaying of an uncompressed DAG 900 shown in FIG. 9 and a compressed DAG 1000 shown in FIG. 10.
[0057] As shown in FIG. 7A, at sub-step 402 of step 304, the data structure can be generated with a backtracking algorithm with use of a ruleset satisfiability oracle. In other words, the computing system can use a ruleset satisfiability oracle and a backtracking algorithm to generate the data structure. As shown in FIG. 7B, at step 304, the data structure can be generated from a Boolean logic formula derived from the given ruleset. In some embodiments, the data structure is generatedfrom a Boolean logic formula derived from the given ruleset with use of an SAT solver or another type of process for solving Boolean satisfiability problems. As shown, in some examples, the method step 304 includes deriving a Boolean logic formula from the given ruleset (at sub-step 502 of step 304, shown in FIG. 7B) and generating the data structure based on the Boolean logic formula (at sub-step 504 of step 304).
[0058] In some embodiments, the method 300 includes the computing system generating a characteristic dependency graph based on the given ruleset (e.g., see step 602 of method 600 shown in FIG. 7C, which can be a part of method 300). In such cases, a feature of the data structure (such as an order of levels of a DAG) is at least partially based on the characteristic dependency graph (e.g., see step or sub-step 604 of step 304, which is shown as being a part of step 304 of method 300). And, in such examples, the method further includes compressing the data structure based on the characteristic dependency graph (e.g., see step 606 of method 600).
[0059] In some embodiments, wherein all valid combinations are known and each of the valid combinations has an associated variable in the LP, and wherein, when all of the valid combinations of a given configuration are represented in the data structure, the computing system uses the data structure in the generation of the columns to solve the LP and to resolve having a very large number of combinations which is associated with the technical problem (e.g., see method steps 802, 804, 806, 808, and 810 of method 800 depicted in FIG. 8, which can be a part or used in combination with method 300). Also, in some embodiments, the computing system determines a suggestion for changing input parameters for the LP in a least intrusive way according to a target function, when the LP has no feasible solution (e.g., see step 812 of method 800 shown in FIG. 8). And, in some examples, the computing system uses historical data to determine a solution that minimizes total deviation from past component consumption when the LP has multiple feasible solutions (e.g., see step 814 of method 800 shown in FIG. 8).
[0060] In some embodiments, the method 300 includes repeating the generation of the data structure and the formulation of the LP, for the given ruleset, multiple times to further enhance a final instance of the data structure outputted by the computing system. In the repeating of the generation of the data structure, the computing system stores each generated instance of the data structure so that multiple data structures are retrievable and determine results of the given ruleset initially before the generation of the multiple instances of the data structure. The computing systemalso can determine a sequence of nodes or other types of features of the multiple instances of the data structure according to repetition in attributes of the results of the given ruleset initially before the generation of the multiple instances of the data structures.
[0061] In some embodiments, the computing system applies the functionality described herein to long-term sales option-based planning. In some examples, the long-term sales option-based planning includes sales option-based planning of configured products. And, in some cases, the configured products include complex models of products.
[0062] In some embodiments, the configured product is a configured agricultural or farming product. And, in some examples, the product is or includes a configured farming machine, a configured agricultural software, a configured hardware system for a farming machine, a configured software system for a farming machine, a configured system including farming machine hardware and software that supports or controls the hardware, or any combination thereof.
[0063] FIG. 7D shows a method 700 which can be a precursor to method 300. Method 700 starts with the computing system using a module of an ERP system to generate the ruleset (at step 702). Then, the method continues with the computing system using the module of the ERP system to generate the valid variants based on the generated ruleset (at step 704). Finally, the method includes using the generated valid variants to generate the structure (at step 706). In some embodiments, the method 300, method 700, or a combination of the two methods is integrated and implemented through the ERP system.
[0064] In some embodiments of the methods described herein, the data structure includes a graph, and the graph includes a DAG or another type of graph. In example embodiments where the data structure is or includes a DAG, the method 300 includes formulating, by the computing system, the LP to solve a technical problem about variants of a configured product (such as an agricultural or farming configured product) permitted by the given ruleset, wherein the LP is formulated with a linear objective function on variables that represent characteristic values of the product and where the method further includes generating, by the computing system, the DAG to determine at least some of the variants of the product. In such examples, the DAG encodes valid variants according to the given ruleset, and each traversal of the DAG from the root node of the DAG to a terminal node of the DAG represents a single valid variant of the configured agricultural product. And, a level ofthe DAG corresponds to a specific characteristic and an edge in a traversal of the DAG corresponds to a choice of a value for an associated characteristic. Further, in such examples, the method includes generating, by the computing system, columns for the LP according to the DAG.
[0065] In some embodiments, the method further includes displaying the DAG in a GUI. For example, see FIG. 9 which illustrates an image of an uncompressed DAG 900 displayed in a part of a GUI. Also, for example, see FIG. 10 which illustrates an image of a compressed DAG 1000 displayed in a part of a GUI, which is at least compressed vertically. And, for example, see FIG. 11 which illustrates images of an uncompressed DAG 1100 and a horizontally compressed DAG 1102 displayed in respective parts of a GUI. Referring back to the method 300, in some examples, the method further includes: storing in memory, by the computing system, the DAG, wherein the DAG is an uncompressed DAG; compressing, by the computing system, the stored DAG; and displaying the uncompressed DAG along with the compressed DAG in the GUI. E.g., see the displaying of the uncompressed DAG 900 shown in FIG. 9 and the compressed DAG 1000 shown in FIG. 10. Also, in some embodiments, the method includes compressing the DAG vertically (e.g., see compressed DAG 1000, which is vertically compressed between “hood” and “engine” levels of the DAG in that, for example, the “frontloader” and “AC” levels are removed in the image). Also, in some embodiments, the method includes compressing the DAG horizontally (e.g., see compressed DAG 1102, which is horizontally compressed at the “color” level of the DAG). Also, in some embodiments, the method includes compressing the DAG vertically and horizontally. Furthermore, in some examples, the method includes selecting, by the computing system, a sub-graph of the DAG and compressing the selected sub-graph instead of the entirety of the DAG.
[0066] Also, similar to the more general examples, when the DAG is used, the solving of the LP requires the generation of the columns in some examples, and the DAG is generated with a backtracking algorithm with use of a ruleset satisfiability oracle in some instances. Also, in some cases, the DAG is generated from a Boolean logic formula derived from the given ruleset, such as with use of an SAT solver. And, in some cases, the generation of the columns is based on Dantzig- Wolfe decomposition, the generation of the columns is based on traversals of the DAG, and the method includes, in solving the LP, delaying column generation to improve tractability of the LP.
[0067] In some examples, the method includes generating a characteristic dependency graph based on the given ruleset and an order of levels of the DAG is at least partially based on thecharacteristic dependency graph. And, in such examples, the method further includes compressing the DAG based on the characteristic dependency graph.
[0068] In some examples, wherein all valid combinations are known and each of the valid combinations has an associated variable in the LP, and wherein, when all of the valid combinations of a given configuration are represented in the DAG, the computing system uses the DAG in the generation of the columns to solve the LP and to resolve having a very large number of combinations which is associated with the technical problem. Also, in some examples using the DAG, the method includes determining, by the computing system, a suggestion for changing input parameters for the LP in a least intrusive way according to a target function, when the LP has no feasible solution. And, in such cases, the method includes using, by the computing system, historical data to determine a solution that minimizes total deviation from past component consumption, when the LP has multiple feasible solutions.
[0069] In some examples, the method includes repeating the generating of the DAG and the formulating of the LP, for the given ruleset, multiple times to further enhance a final DAG outputted by the method. And, the method can include, in the repeating of the generating of the DAG: storing each generated instance of the DAG so that multiple DAGs are retrievable; determining results of the given ruleset initially before the generation of the multiple DAGs; and determining a sequence of nodes of the multiple DAGs according to repetition in attributes of the results of the given ruleset initially before the generation of the multiple DAGs.
[0070] In some embodiments using the DAG, the method includes using a module of an ERP system to generate the ruleset as well as using the module of the ERP system to generate the valid variants based on the generated ruleset and using the generated valid variants to generate the DAG. In some of such examples, the method is integrated and implemented through the ERP system.
[0071] Some portions of the preceding detailed descriptions have been presented in terms of algorithms and symbolic representations of operations on data bits within a computer memory. These algorithmic descriptions and representations are the ways used by those skilled in the data processing arts to most effectively convey the substance of their work to others skilled in the art. An algorithm is here, and generally, conceived to be a self-consistent sequence of operations leading to a predetermined result. The operations are those requiring physical manipulations of physicalquantities. Usually, though not necessarily, these quantities take the form of electrical or magnetic signals capable of being stored, combined, compared, and otherwise manipulated. It has proven convenient at times, principally for reasons of common usage, to refer to these signals as bits, values, elements, symbols, characters, terms, numbers, or the like. It should be borne in mind, however, that these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities. The present disclosure can refer to the action and processes of a computing system, or similar electronic computing device, which manipulates and transforms data represented as physical (electronic) quantities within the computing system's registers and memories into other data similarly represented as physical quantities within the computing system memories or registers or other such information storage systems.
[0072] The present disclosure also relates to an apparatus for performing the operations herein. This apparatus can be specially constructed for the intended purposes, or it can include a general- purpose computer selectively activated or reconfigured by a computer program stored in the computer. Such a computer program can be stored in a computer-readable storage medium, such as any type of disk including floppy disks, optical disks, CD-ROMs, and magnetic-optical disks, readonly memories (ROMs), random access memories (RAMs), EPROMs, EEPROMs, magnetic or optical cards, or any type of media suitable for storing electronic instructions, each coupled to a computing system bus.
[0073] The algorithms and displays presented herein are not inherently related to any particular computer or other apparatus. Various general-purpose systems can be used with programs in accordance with the teachings herein, or it can prove convenient to construct a more specialized apparatus to perform the methods. The structure for a variety of these systems will appear as set forth in the description herein. In addition, the present disclosure is not described with reference to any particular programming language. It will be appreciated that a variety of programming languages can be used to implement the teachings of the disclosure as described herein.
[0074] The present disclosure can be provided as a computer program product, or software, which can include a machine-readable medium having stored thereon instructions, which can be used to program a computing system (or other electronic devices) to perform a process according to the present disclosure. A machine-readable medium includes any mechanism for storing information in a form readable by a machine (e.g., a computer). In some embodiments, a machine-readable (e.g., 1computer-readable) medium includes a machine (e.g., a computer) readable storage medium such as a read only memory (“ROM”), random access memory (“RAM”), magnetic disk storage media, optical storage media, flash memory components, etc.
[0075] While the invention has been described in conjunction with the specific embodiments described herein, it is evident that many alternatives, combinations, modifications and variations are apparent to those skilled in the art. Accordingly, the example embodiments of the invention, as set forth herein are intended to be illustrative only, and not in a limiting sense. Various changes can be made without departing from the spirit and scope of the invention.
Claims
What is claimed is:
1. A method, comprising: formulating, by a computing system, a linear program (LP) to solve a technical problem about variants of a configured agricultural product permitted by a given ruleset, wherein the LP is formulated with a linear objective function on variables that represent characteristic values of the product; generating, by the computing system, a directed acyclic graph (DAG) to determine at least some of the variants of the product, wherein the DAG encodes valid variants according to the given ruleset, wherein each traversal of the DAG from the root node of the DAG to a terminal node of the DAG represents a single valid variant of the configured agricultural product, and wherein a level of the DAG corresponds to a specific characteristic and an edge in a traversal of the DAG corresponds to a choice of a value for an associated characteristic; and generating, by the computing system, columns for the LP according to the DAG.
2. The method of claim 1, further comprising displaying the DAG in a graphical user interface (GUI).
3. The method of claim 2, further comprising: storing in memory, by the computing system, the DAG, wherein the DAG is an uncompressed DAG; compressing, by the computing system, the stored DAG; and displaying the uncompressed DAG along with the compressed DAG in the GUI.
4. The method of claim 1 , wherein solving the LP requires the generation of the columns.
5. The method of claim 1, wherein the DAG is generated with a backtracking algorithm with use of a ruleset satisfiability oracle.
6. The method of claim 1 , wherein the DAG is generated from a Boolean logic formula derived from the given ruleset.
7. The method of claim 1 , wherein the generation of the columns is based on Dantzig-Wolfe decomposition, wherein the generation of the columns is based on traversals of the DAG, and wherein the method comprises, in solving the LP, delaying column generation to improve tractability of the LP.
8. The method of claim 1, wherein the method comprises generating a characteristic dependency graph based on the given ruleset, wherein an order of levels of the DAG is at least partially based on the characteristic dependency graph, and wherein the method further comprises compressing the DAG based on the characteristic dependency graph.
9. The method of claim 1 , wherein all valid combinations are known and each of the valid combinations has an associated variable in the LP, and wherein, when all of the valid combinations of a given configuration are represented in the DAG, the computing system uses the DAG in the generation of the columns to solve the LP and to resolve having a very large number of combinations which is associated with the technical problem.
10. The method of claim 1, comprising: determining, by the computing system, a suggestion for changing input parameters for the LP in a least intrusive way according to a target function, when the LP has no feasible solution; and using, by the computing system, historical data to determine a solution that minimizes total deviation from past component consumption, when the LP has multiple feasible solutions.
11. The method of claim 1 , comprising repeating the generating of the DAG and the formulating of the LP, for the given ruleset, multiple times to further enhance a final DAG outputted by the method.
12. The method of claim 11, comprising, in the repeating of the generating of the DAG: storing each generated instance of the DAG so that multiple DAGs are retrievable; determining results of the given ruleset initially, before the generation of the multiple DAGs; and determining a sequence of nodes of the multiple DAGs according to repetition in attributes of the results of the given ruleset initially before the generation of the multiple DAGs.
13. The method of claim 1, wherein the configured agricultural product comprises: a configured farming machine, configured agricultural software, a configured hardware system for a farming machine, a configured software system for a farming machine, a configured system comprising farming machine hardware and software that supports or controls the hardware, or any combination thereof.
14. The method of claim 1, wherein the method is for long-term sales option-based planning, wherein the long-term sales option-based planning includes sales option-based planning of configured agriculture products, and wherein the configured agriculture products include complex models of agricultural products.
15. The method of claim 1, comprising: using a module of an enterprise resource planning (ERP) system to generate the ruleset; using the module of the ERP system to generate the valid variants based on the generated ruleset; and using the generated valid variants to generate the DAG.
16. The method of claim 15, wherein the method is integrated and implemented through the ERP system.
17. The method of claim 1, comprising: compressing the DAG vertically, horizontally, or any combination thereof; and selecting a sub-graph of the DAG and compressing the selected sub-graph.
18. A method, comprising: formulating, by a computing system, a linear program (LP) to solve a technical problem about variants of a configured agricultural product permitted by a given ruleset, wherein the LP is formulated with a linear objective function on variables that represent characteristic values of the product; generating, by the computing system, a data structure to determine at least some of the variants of the product, wherein the data structure encodes valid variants according to the given ruleset, and wherein each iterative use of the data structure represents a single valid variant of the configured agricultural product; andgenerating, by the computing system, columns for the LP according to the data structure.
19. The method of claim 18, wherein the data structure comprises a graph, and wherein the graph comprises a directed acyclic graph (DAG) or another type of graph.
20. A computer system, comprising: a processing device; and memory in communication with the processing device and storing instructions that, when executed by the processing device, cause the processing device to: formulate a linear program (LP) to solve a technical problem about variants of a configured agricultural product permitted by a given ruleset, wherein the LP is formulated with a linear objective function on variables that represent characteristic values of the product; generate a data structure to determine at least some of the variants of the product, wherein the data structure encodes valid variants according to the given ruleset, and wherein each iterative use of the data structure represents a single valid variant of the configured agricultural product; and generate columns for the LP according to the data structure.
Citation Information
Patent Citations
Column generation methods and systems for routing and scheduling in robotic and vehicular applications
US20230075128A1
