Productivity management system

Abstract

The invention discloses a capacity management system. The capacity management system comprises a receiving unit, a calculating unit, a comparing unit and an output unit, the receiving unit is used for receiving a plurality of production targets, the capacity supply quantity of the first production machine and the capacity supply quantity of the second production machine. The calculation unit is used for calculating a plurality of arrangement total scores of the plurality of production marks distributed to the first production machine and the second production machine in different numbers. The comparison unit is used for comparing the plurality of arrangement total scores and selecting the maximum arrangement total score. The output unit is used for outputting the number of the plurality of production targets distributed to the first production machine and the second production machine with the maximum ranking total score.

Description

TECHNICAL FIELD

[0001] The present invention relates to a capacity management system, and in particular, to a capacity management system for maximizing capacity according to multiple key performance indicators (KPIs). BACKGROUND

[0002] Manufacturing industry is the backbone of modern economic system. With the development of global economy, cross-border and cross-region production is becoming more and more common. Whether it is the semiconductor manufacturing and / or packaging industry or the display industry, it is constantly facing increasingly complex and competitive challenges. In this changing and complex environment, for example in the field of thin film transistor liquid crystal display (TFT-LCD), the accuracy and efficiency of production planning directly affect the competitiveness of manufacturers. The manufacturing and assembly of liquid crystal display modules (LCM) is a key link in the entire production process. However, in the face of rapid changes in market demand, the complexity of the supply chain and the dispersion of production facilities, capacity allocation and production planning become increasingly difficult. At the same time, the influence of war and / or disease, global economic environment and the business layout of competitors often affect the order changes of a company. In response to these changes, business strategies must be constantly adjusted, such as: adjusting the load reduction of TFT-LCD, thereby reducing the input allocation of glass, which may result in a reduction in input planning. Therefore, manufacturers not only need to respond to market changes, but also need to flexibly adjust production planning to ensure that production planning meets the principle of "cost reduction and efficiency improvement", so that manufacturers can maintain competitiveness in a competitive environment. SUMMARY

[0003] The purpose of the present invention is to provide a capacity management system.

[0004] The present invention provides a capacity management system, which includes a receiving unit, a calculating unit, a comparing unit and an output unit. The receiving unit is used to receive multiple production targets and the capacity supply amount of a first production machine and the capacity supply amount of a second production machine. The calculating unit is used to calculate multiple allocation total scores of the multiple production targets allocated to the first production machine and the second production machine in different quantities. The comparing unit is used to compare the multiple allocation total scores and select the maximum allocation total score. The output unit is used to output the quantity of the multiple production targets allocated to the first production machine and the second production machine respectively by the maximum allocation total score. BRIEF DESCRIPTION OF DRAWINGS

[0005] Figure 1 is a functional block diagram of a capacity management system according to an embodiment of the present invention coupled to a cloud system.

[0006] Figure 2 is Figure 1 the capacity management system of

[0007] Figure 3 is a flowchart of the step of intelligent engine operation in Figure 2

[0008] BRIEF DESCRIPTION OF DRAWINGS100 - capacity management system; 101 - receiving unit; 102 - calculating unit; 103 - comparing unit; 104 - output unit; 105 - storing unit; 106 - network unit; 107 - display unit; 120 - input device; 130 - network; 150 - cloud system; 200 - method; 300 - total score; S210, S220, S222, S224, S226, S230 - steps. DETAILED DESCRIPTION

[0009] The present application can be understood with reference to the following detailed description and drawings in which:

[0010] Certain terms are used throughout the present application and the claims to refer to particular components. As one skilled in the art will appreciate, electronic equipment manufacturers can refer to a component by different names. This document does not intend to distinguish between components that differ in name but not in function. In the description and claims of the application, each of the words "comprise", "comprises", "comprising" and the like are to be read expansively and without limitation. Thus, these terms include the meaning that "comprising" or "comprises" means "including, but not limited to".

[0011] In the description and claims of the application, each of the words "comprise", "comprises", "comprising" and the like are to be read expansively and without limitation. Thus, these terms include the meaning that "comprising" or "comprises" means "including, but not limited to".

[0012] The directional terms used in the present application, such as "upper", "lower", "front", "back", "left", "right", and the like, are only with reference to the direction of the drawings. Therefore, the directional terms used are for explanation, and not for limiting the present application. In the drawings, each drawing shows the general features of the methods, structures and / or materials used in a particular embodiment. However, these drawings should not be interpreted as defining or limiting the scope or nature of what is encompassed by these embodiments. For example, the relative sizes, thicknesses and positions of each film layer, region and / or structure can be reduced or enlarged for clarity.

[0013] When a component (e.g., a membrane or region) is referred to as "on another component," it can be directly on that component, or there may be other components between them. Conversely, when a component is referred to as "directly on another component," there are no components between them. Furthermore, when a component is referred to as "on another component," the two are vertically related, and this component can be above or below the other component, depending on the orientation of the device.

[0014] It should be understood that when a component or membrane is referred to as being "connected to" another component or membrane, it can be directly connected to this other component or membrane, or there can be an intercalated component or membrane between them. When a component is referred to as being "directly connected to" another component or membrane, there is no intercalated component or membrane between them. Additionally, when a component is referred to as being "coupled to another component (or a variant thereof)," it can be directly electrically connected to this other component, or indirectly connected (e.g., indirectly electrically connected) to this other component through one or more components.

[0015] In this invention, when one component "disconnects" from another component, an electrical signal cannot flow between the two components for a specified period of time.

[0016] The terms “approximately” or “about” are generally interpreted as being within ±10% of a given value, or within ±5%, ±3%, ±2%, ±1%, or ±0.5% of a given value.

[0017] The ordinal numbers used in the specification and claims of this invention, such as "first," "second," etc., to modify elements, do not in themselves imply or represent any prior ordinal number of that element (or those elements), nor do they represent the order of one element with another, or the order of manufacturing methods. The use of these ordinal numbers is solely to clearly distinguish one named element from another element with the same name. The claims and specification may not use the same terminology; therefore, a first element in the specification may be a second element in the claims.

[0018] It should be understood that the features described in the following embodiments can be replaced, recombined, or mixed in several different embodiments to complete other embodiments without departing from the spirit of the invention. Features between embodiments can be arbitrarily mixed and combined as long as they do not violate the spirit of the invention or conflict with it.

[0019] In this invention, the electronic device may include a display device, a light-emitting device, an antenna device, a sensing device, a splicing device, or any combination thereof, but is not limited thereto. The display device may be a non-self-emissive display or a self-emissive display, and may be a color display or a monochrome display, depending on the requirements. The antenna device may be a liquid crystal type antenna device or a non-liquid crystal type antenna device; the sensing device may be a sensing device for capacitance, light, heat, or ultrasound; and the splicing device may be a display splicing device or an antenna splicing device, but is not limited thereto. The electronic device may include electronic components, which may include passive and active components, such as capacitors, resistors, inductors, diodes, transistors, dies, or chips. Diodes may be dies or chips, and may include light-emitting diodes (LEDs), photodiodes, or varactors, but are not limited thereto. Light-emitting diodes (LEDs) may include, for example, organic light-emitting diodes (OLEDs), mini LEDs, micro LEDs, or quantum dot LEDs, but are not limited thereto. Transistors may include, for example, top-gate thin-film transistors, bottom-gate thin-film transistors, or dual-gate thin-film transistors, but are not limited thereto. Electronic devices may also include fluorescent materials, phosphorescent materials, quantum dot (QD) materials, or other suitable materials as needed, but are not limited thereto. Electronic devices may have peripheral systems such as drive systems, control systems, light source systems, etc., to support the devices and components within the electronic device.

[0020] It should be noted that the technical features in the different embodiments described below may be replaced, reorganized or combined with each other to form another embodiment without departing from the spirit of the invention.

[0021] Please refer to Figure 1 , Figure 1This is a functional block diagram of a capacity management system 100 according to an embodiment of the present invention, which can be connected to a cloud system 150 via a network 130. The capacity management system 100 is used to allocate the production of multiple production targets to multiple production machines, so that the multiple production machines manufacture the production targets according to the allocated production quantities. The aforementioned production targets are products (e.g., electronic devices) that can be manufactured by the production machines. The aforementioned multiple production machines can be configured in the same factory area or in multiple different factory areas. The capacity management system 100 may include a receiving unit 101, a calculation unit 102, a comparison unit 103, and an output unit 104. The receiving unit 101 is used to receive data 160. Data 160 includes, but is not limited to, the capacity supply of multiple production targets and multiple production machines. For example, in one embodiment of the present invention, data 160 records the capacity supply of three production machines for a certain production target among multiple production targets in different months, which can be represented in Table 1 below:

[0022] (Table 1)

[0023]

[0024] Here, 1K represents one thousand, while 10K represents ten thousand, and so on.

[0025] In one embodiment of the present invention, the receiving unit 101 may receive data 160 from the cloud system 150 via the network 130. In another embodiment, the receiving unit 101 may receive data 160 via the input device 120, but the present invention is not limited thereto.

[0026] The calculation unit 102 is used to calculate multiple production targets, allocating them in different quantities to multiple production machines, and calculating multiple allocation totals based on data 160. For example, based on the capacity supply in January, it allocates different quantities to the three production machines in Table 1, finds the optimal allocation total, and uses it as a reference for allocating the capacity supply in February. Table 2 below shows the multiple allocation totals calculated by the calculation unit 102 based on the capacity supply in January from data 160, allocating different quantities to the three production machines in Table 1.

[0027] (Table 2)

[0028]

[0029] The comparison unit 103 compares the total allocation scores calculated by the calculation unit 102 and selects the maximum value. Taking Table 2 as an example, the comparison unit 103 compares the total allocation scores of the three allocation scenarios in Table 2 and obtains the maximum total allocation score obtained in allocation scenario one. Then, the output unit 104 outputs the quantity of each production machine corresponding to the production target in Table 1 in scenario one, that is: the production machines numbered 1, 2 and 3 are allocated 700K, 8500K and 1000K respectively.

[0030] The receiving unit 101 may be, but is not limited to, a Universal Serial Bus (USB), a Peripheral Component Interconnect Express (PCI-E), an RJ45 network interface, a fiber optic network interface, a coaxial cable network interface, a modem, a Wi-Fi module, a Bluetooth module, a Zigbee module, a WiMAX module, a fourth-generation (4G) mobile communication module, or a fifth-generation (5G) mobile communication module. The computing unit 102 and the comparison unit 103 may each be an integrated circuit (IC), such as a processor or a central processing unit (CPU), but the invention is not limited thereto. In another embodiment, the computing unit 102 and the comparison unit 103 may be integrated into the same integrated circuit (e.g., the same processor). Output unit 104 may be, but is not limited to, a Universal Serial Bus (USB), a Peripheral Component Interconnect Express (PCI-E), a graphics card interface (such as VGA, DVI, HDMI, or DisplayPort), and a printer interface.

[0031] In one embodiment of the present invention, the capacity management system 100 may also be coupled to the display unit 107 via wired or wireless means to display the quantity of each production target allocated to each production machine output by the output unit 104. The display unit 107 may include a display device according to usage requirements, including but not limited to a non-emissive display or a self-emissive display, and may be a color display or a monochrome display according to usage requirements.

[0032] In one embodiment of the present invention, the capacity management system 100 can also receive data and / or instructions input by the user through the input device 120 from the receiving unit 101, so that the capacity management system 100 can perform corresponding operations based on the data and / or instructions input by the user through the input device 120. The input device 120 may be, but is not limited to, a keyboard, a mouse, or a barcode scanner.

[0033] In one embodiment of the present invention, the capacity management system 100 may further include a storage unit 105 for storing programs to be executed by the capacity management system 100, data 160 received by the receiving unit 101, data generated by the computing unit 102 during the computing process, and / or data output by the output unit 104. The storage unit 105 may be, but is not limited to, dynamic random access memory, static random access memory, flash memory, floppy disk, hard disk, optical disk, USB flash drive, magnetic tape, or a combination thereof. In one embodiment, the computing unit 102 may access and execute the programs stored in the storage unit 105 to achieve the functions to be performed by the capacity management system 100.

[0034] In one embodiment of the present invention, the capacity management system 100 may further include a network unit 106, which can be coupled to a cloud system 150 via a network 130. The network 130 may be a local area network or the Internet. The cloud system 150 may be a server or another capacity management system 100. In one embodiment, the receiving unit 101 can be coupled to the cloud system 150 via the network unit 106 and the network 130 and receive data 160 required by the computing unit 102.

[0035] The Capacity Management System 100 introduces a Key Performance Indicator (KPI) adjustment mechanism, and incorporates mathematical models and optimization algorithms to further adjust the KPIs. Simultaneously, the Capacity Management System 100 can integrate information from inter-plant connections to strengthen scheduling monitoring and reporting systems, improve scheduling visibility and / or transparency, and more effectively respond to production changes. Please refer to [reference needed]. Figure 2 , Figure 2 yes Figure 1 The flowchart illustrates the execution of capacity management method 200 by the capacity management system 100. Management method 200 includes the following steps:

[0036] Step S210: System serial connection;

[0037] Step S220: Intelligent engine calculation; and

[0038] Step S230: Visualize decision support.

[0039] Step S210 mainly involves integrating and organizing the data 160 from the systems of various factories to facilitate use and calculation by the calculation unit 102. Step S220 mainly includes three steps: "KPI logic adjustment procedure", "KPI weight optimization procedure" and "multi-objective optimization procedure", which will be further explained below. Step S230 involves presenting the quantities of each production target assigned to each production machine output by the comparison unit 103 in a visual manner (e.g., through the display unit 107).

[0040] Please refer to Figure 3 , Figure 3 Is execution Figure 2 A flowchart for step S220. Step S220 includes:

[0041] Step S222: The calculation unit 102 executes the KPI logic adjustment program based on the data 160;

[0042] Step S224: The calculation unit 102 executes the KPI weight optimization procedure based on the data 160; and

[0043] Step S226: The calculation unit 102 executes a multi-objective optimization program to calculate the quantity of each production target allocated to each production machine. Steps S222, S224, and S226 will be explained further below.

[0044] The following is for now. Figure 3 The "Step S222: KPI Logic Adjustment Procedure" section explains this process. Step S222 primarily involves comprehensively considering the capacity of production machines, the productivity of each plant area, capacity balance, and / or the efficiency of automated production lines when seeking the optimal scheduling scheme, in order to maximize production efficiency. This comprehensive consideration ensures that the scheduling scheme takes into account production needs, equipment efficiency, and overall benefits. Through the systematic selection and comprehensive consideration by the capacity management system 100, production managers can gain a more comprehensive understanding of the production environment, fully utilize domain expert knowledge, incorporate these key factors into the calculation logic, and embed multi-objective optimization algorithms to achieve more accurate and efficient automated scheduling. The capacity management system 100 not only ensures that production needs are met but also enables the scheduling scheme to achieve optimal performance in terms of equipment efficiency and overall benefits. In step S222, the calculation unit 102 of the capacity management system 100 defines key performance indicators (KPIs), and the KPIs used by the calculation unit 102 may include, but are not limited to, the following four types: KPI1 (production cost variation rate), KPI2 (scarce equipment capacity), KPI3 (automated production line utilization rate) and KPI4 (balance between plants).

[0045] (Table 3)

[0046]

[0047] Table 3 above shows the target values ​​and adjustment directions for each KPI. Specifically, for "Production Cost Variation Rate," calculation unit 102 uses the comparison between actual production cost and theoretical optimal cost as the target value, aiming for costs to approach or near the optimal theoretical cost to achieve more effective cost control; its adjustment direction is "target." Secondly, regarding "Scarce Equipment Capacity," calculation unit 102 aims for lower loading rates to ensure full capacity utilization, keeping the utilization rate of scarce equipment below a preset percentage (e.g., 90%) to prevent potential shortages. For "Automated Production Line Utilization Rate," calculation unit 102 emphasizes maximizing the utilization of automated production lines to improve efficiency and reduce production costs. Regarding "Plant Area Balance," calculation unit 102 primarily aims to bring actual allocations closer to target allocations to ensure the full utilization of production capacity in each plant area; its adjustment direction is "target." These KPI settings are designed to assist the capacity management system 100 in achieving comprehensive production optimization and efficiency improvement.

[0048] In step S222, after the calculation unit 102 completes the definition of each KPI, it can further standardize each KPI. Taking KPI1 (production cost variation rate) as an example, the calculation unit 102 will calculate the target value of production cost (i.e., achievement rate). The achievement rate is defined as shown in Table 3 as equal to (production cost after allocation) divided by (theoretical optimal cost). The production cost of allocation must be greater than or equal to the theoretical optimal cost, so the achievement rate must be greater than or equal to 1. Now there are three products to be allocated, and the connected plants and corresponding costs for each product are shown in Table 4 below:

[0049] (Table 4)

[0050] Production target Connected plants Single production cost A Plant 3 120 A Plant 1 270 A Plant 4 90 B Plant 2 200 B Plant 3 100 C Plant 1 350 C Plant 4 200

[0051] If each of the three production targets (i.e., products) produces only one unit, then the theoretically optimal cost (product produced in the lowest-cost plant) is (90 + 100 + 200), which equals 390. Assuming production targets A, B, and C are allocated to plant 3, plant 4, and plant 4 respectively, their production cost after allocation is (120 + 100 + 200), which equals 420. Therefore, the achievement rate is 420 divided by 390, which equals 1.08. Then, if (1 - abs(1 - achievement rate)) is less than 0, the standardized score of KPI1 is 100; if (1 - abs(1 - achievement rate)) is greater than 0, the standardized score of KPI1 is ([1 - abs(1 - achievement rate)] × 100). Taking the above as an example, if the achievement rate is 1.08, then the standardized score of its KPI1 is ([1-abs(1-1.08)]×100)=92.

[0052] When standardizing KPI2 (equipment capacity load rate), calculation unit 102 first calculates its load rate (load rate = allocation demand / capacity supply) according to Table 3, and then standardizes KPI2 based on the load rate. Specifically, when the load rate ≤ N, the score corresponding to KPI2 is 100; when the load rate > 1, the score corresponding to KPI2 is (N - load rate) × 100; when N ≤ load rate ≤ 1, the score corresponding to KPI2 is (1 - (load rate - N))) × 100; if capacity supply = 0 and allocation demand > 0, then the load rate is set to 9.99, and the score corresponding to KPI2 is (1 - (9.99 - N)) * 100. N can be 0.95, 0.9, 0.85, or 0.8, and can be selected according to the situation. In one embodiment, for example, N = 0.9. Assume that the allocation demand and capacity supply of a certain production target (i.e., product) are represented by Tables 5 and 6 below:

[0053] (Table 5)

[0054]

[0055] (Table 6)

[0056]

[0057] The load rates calculated by calculation unit 102 based on Tables 5 and 6 are shown in Table 7 below:

[0058] (Table 7)

[0059]

[0060] Calculation unit 102 converts Table 7 into the scores corresponding to KPI2 as shown in Table 8 below:

[0061] (Table 8)

[0062] Plant Score for January 2024 Score for February 2024 Score for March 2024 1 100 100 -9.09 2 100 -16 -43 3 100 100 100

[0063] In addition, the calculation unit 102 will convert Table 5 into the following Table 9:

[0064] (Table 9)

[0065]

[0066] Then, the calculation unit 102 multiplies the values ​​in Table 5 and Table 9 to obtain the following table 10:

[0067] (Table 10)

[0068]

[0069] KPI2 (equipment capacity utilization) can, for example, cover the following five sub-indicators: FHD, THICKNESS, Aspect Ratio, TOUCH, and UHD / QHD. FHD represents the bonding equipment capacity required for high-resolution (1920x1080) products; THICKNESS represents the bonding equipment capacity required for products with a substrate (e.g., glass) thickness ≤ 0.25 cm; TOUCH represents the bonding equipment capacity required for products with touch functionality; Aspect Ratio represents the bonding equipment capacity required for products with special dimensions, such as those used in certain notebook computer screens (NB) with a 16:10 aspect ratio; UHD / QHD represents the bonding equipment capacity required for products used as displays (MNTs) and are of Ultra High Definition (UHD, resolution 3840x2160) or Quarter High Definition (QHD, resolution 2560x1440). Tables 5 to 10 above show the standardized KPI2 score for one of the five sub-item indicators. The other four sub-item indicators can be standardized in a similar manner. Finally, the calculation unit 102 averages the scores of the five sub-item indicators to obtain the standardized KPI2 score.

[0070] When standardizing KPI3 (automated production line utilization rate), calculation unit 102 first calculates its utilization rate (utilization rate = allocation demand / capacity supply) according to Table 3, and then standardizes KPI3 based on the utilization rate. Specifically, when the utilization rate ≥ 1, the score corresponding to KPI3 is 100; when the utilization rate < 1, the score corresponding to KPI3 is (utilization rate × 100). Assume that the allocation demand and capacity supply of a certain production target (i.e., product) are represented by Tables 11 and 12 below:

[0071] (Table 11)

[0072]

[0073] (Table 12)

[0074]

[0075] The utilization rate calculated by calculation unit 102 based on Tables 11 and 12 is shown in Table 13 below:

[0076] (Table 13)

[0077]

[0078] Calculation unit 102 converts Table 13 into the scores corresponding to KPI3 as shown in Table 14 below:

[0079] (Table 14)

[0080] Plant Score for January 2024 Score for February 2024 Score for March 2024 1 8 47 11 2 100 100 100 3 0 0 0

[0081] In addition, the calculation unit 102 will convert Table 12 into the following Table 15:

[0082] (Table 15)

[0083]

[0084] Then, the calculation unit 102 multiplies the values ​​in Table 11 and Table 15 to obtain the following table 16:

[0085] (Table 16)

[0086]

[0087] KPI3 can cover two sub-indicators: notebook computers (NB) and monitors (MNT). Tables 11 to 16 above show the standardized KPI3 score for one of these two sub-indicators. The other sub-indicator can be standardized in a similar manner. Finally, the calculation unit 102 averages the scores of the two sub-indicators to obtain the standardized KPI3 score.

[0088] When standardizing KPI4 (balance across plant areas), calculation unit 102 first calculates its achievement rate (achievement rate = actual allocation / target allocation) based on Table 3, and then standardizes KPI4 based on the achievement rate. The score corresponding to KPI4 is equal to ([1 - abs(1 - achievement rate)] × 100). Assuming that calculation unit 102 provides a target allocation ratio for each allocation as the target, and then calculates the achievement rate with the actual allocation, Table 17 below shows the target ratio of products in each plant area when products are allocated to plants A and B:

[0089] (Table 17)

[0090]

[0091] If the total production to be allocated each month is as shown in Table 18 below:

[0092] (Table 18)

[0093] January 2024 February 2024 March 2024 250000 300000 500000

[0094] Calculation unit 102 obtains the following table 19 based on Tables 17 and 18:

[0095] (Table 19)

[0096]

[0097] If the actual dosage is as shown in Table 20 below:

[0098] (Table 20)

[0099]

[0100] Calculation unit 102 calculates the achievement rate of KPI4 based on Table 3 (achievement rate = actual allocation / target allocation), resulting in the following table 21:

[0101] (Table 21)

[0102]

[0103] in,

[0104]

[0105] By standardizing the concept, it is ensured that all sub-objectives have a consistent scale in the converted scores, thus resolving the different levels of "ambition," "small-scale," and "large-scale" distinctions in the original KPI target directions. Through the process of converting KPIs into scores, the capacity management system 100 uniformly adjusts the original directions of "ambition," "small-scale," and "large-scale," aiming for all four KPI scores to trend towards improvement, transforming the original Table 3 into the following Table 22. This transformation ensures that the relative importance and achievement of each sub-objective are presented more fairly in the final evaluation. Through this integration, the capacity management system 100 ensures the consistency of KPI evaluation, making the performance of each sub-objective more objective and comparable.

[0106] (Table 22)

[0107]

[0108] The following is for now. Figure 3The following section explains "Step S224: KPI Weight Optimization". Step S224 primarily aims to maximize the total allocation score of 300 by finding a balance between production costs, scarce equipment capacity, automated production line utilization rate, and the balance score of each plant. However, the four KPIs—KPI1 (production costs), KPI2 (scarce equipment capacity), KPI3 (automated line utilization rate), and KPI4 (plant balance)—interfere with each other, making it difficult to achieve a balance. Overemphasizing one KPI inevitably suppresses the others. To address this challenge, the capacity management system 100 defines KPIs to establish a quantitative scoring logic. This logic not only allows production managers to better understand the interrelationships between KPIs through the capacity management system 100 but also enables flexible adjustment of KPI weights in response to changes in market conditions and company policies. Specifically, the calculation unit 102 can calculate the total allocation score of 300 using the following formula:

[0109]

[0110] w represents the weight, KPI represents the key performance indicator, and i represents a variable from 1 to 4.

[0111] In another embodiment of the invention, a total score of 300 is equal to... That is, the total allocation score of 300 is equal to (w1×KPI1+w2×KPI2), where KPI1 is the production cost variation rate, KPI2 is the equipment capacity load rate, and w1 and w2 are the weights corresponding to the production cost variation rate KPI1 and the equipment capacity load rate KPI2, respectively.

[0112] Through step S224, the calculation unit 102 can obtain the optimized weight w corresponding to each KPI. The methods for optimizing KPI weights can be broadly categorized into subjective weighting methods and objective weighting methods. Subjective weighting methods rely heavily on the subjective judgment and determination of weights by individuals or experts, typically based on their experience, knowledge, and intuition to set the weights of each KPI. Subjective weighting methods may include empirical methods, the Analytic Hierarchy Process (AHP), and the Decision-Making Trial and Evaluation Laboratory (DEMATEL) method.

[0113] Objective weighting methods primarily use statistical analysis or mathematical models to calculate weights, thereby reducing subjective bias and improving the objectivity of the analysis. Objective weighting methods include the entropy weight method and principal component analysis (PCA).

[0114] The expert experience method can quickly assign weights by referencing market conditions and company policies. This method is intuitive and adaptable, and can rapidly respond to changes in internal company culture and external factors. For example, when focusing on production costs, the weights of each KPI can be set as shown in Table 23 below:

[0115] (Table 23)

[0116] KPI Weight (%) 1. Production cost variability rate 40 2. Equipment capacity load rate 20 3. Automation line uptime rate 25 4. Plant balance 15 Total 100

[0117] When focusing on production costs and automated line utilization rates, the weights of each KPI can be set as shown in Table 24 below:

[0118] (Table 24)

[0119] KPI Weight (%) 1. Production cost variability rate 35 2. Equipment capacity load rate 20 3. Automation line uptime rate 35 4. Plant balance 10 Total 100

[0120] When focusing on the production capacity of scarce equipment without considering the balance between different factories, the weights of each KPI can be set as shown in Table 25 below:

[0121] (Table 25)

[0122] KPI Weight (%) 1. Production cost variability rate 30 2. Equipment capacity load rate 45 3. Automation line uptime rate 20 4. Plant balance 5 Total 100

[0123] The Analytic Hierarchy Process (AHP) can be used in the following scenarios: setting priorities, generating a set of alternatives, choosing the best policy alternatives, determining requirements, making decisions using benefits and costs, allocating resources, predicting outcomes and risk assessment, measuring performance, designing a system, ensuring system stability, optimization, planning, and conflict resolution. The AHP can establish an evaluation criterion comparison matrix A based on expert experience, and the relative importance of two different scenarios can be quantified according to Table 26 below:

[0124] (Table 26)

[0125] Factor i vs. factor j Quantified value Equal importance 1 Slightly more important 3 More important 5 Strongly important 7 Extremely important 9 Intermediate value between two adjacent judgments 2, 4, 6 or 8

[0126] The Analytic Hierarchy Process (AHP) is an objective and structured method based on mathematical models, primarily applied to decision-making problems involving uncertainty and multiple evaluation criteria. AHP calculates the weight of each KPI through comparison matrices and consistency checks, ensuring high reliability and consistency. Below, A represents the evaluation criterion comparison matrix.

[0127] AW=λW

[0128] Where W is a non-zero vector, and represents the eigenvectors of matrix A;

[0129] λ is the eigenvalue of matrix A.

[0130] In one embodiment, it is assumed that the comparison matrix A can be obtained from the following Table 27:

[0131] (Table 27)

[0132]

[0133] The comparison matrix A can then be represented as shown in Table 28 below:

[0134] (Table 28)

[0135]

[0136] The calculation unit 102 divides each column of Table 28 by its corresponding sum, resulting in Table 29 below:

[0137] (Table 29)

[0138]

[0139] Next, the calculation unit 102 averages the columns of each row in Table 29 to obtain the following table 30:

[0140] (Table 30)

[0141]

[0142] The matrix of weights W is...

[0143] Next, the calculation unit 102 performs matrix operations on Table 27 and the weights W to obtain matrix B. Next, the calculation unit 102 divides matrix B by the matrix of weights W to obtain matrix C, where matrix C is... Calculation unit 102 calculates λmax based on matrix C, where λmax is equal to the average of the values ​​in matrix C, i.e., ((4.354966+4.288506+4.028138+4.2220008)÷4)=4.222905. Calculation unit 102 then calculates the consistency index (CI) and the consistency ratio (CR). The consistency ratio (CR) is calculated as CI / RI. Taking the above example, the consistency index (CI) is 0.074302, and the consistency ratio (CR) is 0.074302 / 0.9 = 0.082558. Here, n is the number of factors. When CR = 0, it indicates complete consistency between the preceding and following judgments. When CR > 0.1, it indicates a discrepancy or inconsistency between the preceding and following judgments. When CR ≤ 0.1, it indicates that while the preceding and following judgments are not completely consistent, the bias is acceptable.

[0144] The Decision Experimentation and Evaluation Laboratory (DEMATEL) method originated from the Battelle Association, a research center in Geneva, in 1973. Its goal is to find solutions to problems such as environmental protection, energy, and race, and it can be used to solve complex and difficult-to-understand problems. The DEMATEL method has a complex causal structure. It compares the relationships between variables, focusing on the degree of pairwise influence between elements, using matrix and mathematical theory to calculate the causal relationships between all elements, and representing the degree of causal influence numerically.

[0145] The Entropy Weight Method, proposed by Shannon in 1948, is a weighting method that calculates the relative weight of evaluation criteria based on the amount of information conveyed by a particular piece of information, thus providing an objective weighting evaluation method. Its advantage lies in the fact that indicators with higher entropy weights represent more important indicators, indicating that they provide more information to decision-makers.

[0146] Principal Component Analysis (PCA) is a statistical method that identifies underlying factors or components driving central changes in data and can be used to determine index weights. Each permutation yields only one set of weights. Assume a permutation result has 5 options, each with 4 KPI scores, as shown in Table 31 below:

[0147] (Table 31)

[0148] KPI1 KPI2 KPI3 KPI4 Scenario 1 10 20 30 40 Scenario 2 15 25 35 45 Scenario 3 20 30 40 50 Scenario 4 25 35 45 55 Scenario 5 30 40 50 60

[0149] When calculation unit 102 performs PCA, it first standardizes the data to ensure that each indicator contributes equally to the analysis. After standardization, the mean of each KPI is 0, and the standard deviation is 1. Table 31, after PCA standardization, becomes Table 32 below:

[0150] (Table 32)

[0151] KPI1 KPI2 KPI3 KPI4 Scenario 1 -1.414 -1.414 -1.414 -1.414 Scenario 2 -0.707 -0.707 -0.707 -0.707 Scenario 3 0 0 0 0 Scenario 4 0.707 0.707 0.707 0.707 Scenario 5 1.414 1.414 1.414 1.414

[0152] Next, the calculation unit 102 calculates the covariance matrix. Where Z is the standardized data matrix, and n is the number of samples (number of arrangement schemes).

[0153]

[0154] After the computation unit 102 performs PCA to obtain the covariance matrix C, the computation unit 102 calculates the eigenvalues ​​λ and eigenvectors v1. The eigenvalues ​​and eigenvectors of the covariance matrix (C) represent the degree of variation of the data in different directions.

[0155]

[0156] Solving the above equations yields the eigenvalues ​​and eigenvectors. Due to the special structure of the covariance matrix, the eigenvalues ​​and eigenvectors are:

[0157] When the eigenvalue λ1 equals 5, the eigenvector

[0158] When the eigenvalue λ2 equals 0, any three mutually orthogonal vectors v2, v3, and v4, and orthogonal to the eigenvector v1:

[0159] vector vector vector

[0160] The calculation unit 102 can calculate the weight of KPI by considering only the first principal component (the eigenvector corresponding to the largest eigenvalue).

[0161] The following is for now. Figure 3 The following section explains "Step S226: Multi-objective Optimization". Step S226 mainly optimizes the production quantity allocated to each factory area when multiple production targets need to be allocated to multiple factory areas for production, thereby maximizing capacity. In other words, before the calculation unit 102 completes the calculation of the total allocation score 300, the calculation unit 102 optimizes the number of production targets allocated to each production machine by executing step S226. Through step S226, the calculation unit 102 performs multi-objective optimization to calculate the number of each production target allocated to each production machine. As mentioned above, multiple production machines can be configured in the same factory area or in multiple different factory areas.

[0162] The computational unit 102 may execute step S226 in the following ways: heuristic algorithms, meta-heuristic algorithms, linear programming, and deep Q-learning, but the present invention is not limited thereto. Heuristic algorithms are problem-solving methods that rely on intuition, experience, or rules of thumb to find solutions. They are typically designed for specific problems and utilize domain-specific knowledge to guide the search for solutions. Compared to heuristic algorithms, meta-heuristic algorithms are higher-level strategies designed to guide the search process across a wide range of problems, while general optimization algorithms provide a flexible framework for solving various problems. Linear programming is used to solve problems involving the minimization or maximization of a single objective function and is a subfield of mathematical optimization. Deep learning is generally applied to single objectives, such as classification or regression.

[0163] Heuristic algorithms can include the Greedy Algorithm (GrA). A greedy algorithm is an algorithm that, at each step, makes the best or optimal choice given the current state, hoping to achieve the best or optimal result. The greedy algorithm finds the best solution based on the "current" state at each step, focusing on which KPI has a higher weight. The following example illustrates a greedy algorithm:

[0164] (Table 33)

[0165]

[0166] When executing the greedy algorithm, the calculation unit 102 first establishes the following two greedy rules: (1) Cost priority order: 10→15→20→30→35→40; and (2) Cost difference priority order: Product 1→(Product 2; Product 3; Product 4)→Product 5. When the calculation unit 102 performs the first calculation of the greedy algorithm, the calculation unit 102 first takes KPI1 (production cost variation rate) as the priority consideration, and selects factory area C to manufacture product 1, factory area A to manufacture product 2, factory area A to manufacture product 3, factory area A to manufacture product 4 and factory area B to manufacture product 5 respectively. Then its total production cost is equal to (15+10+20+30+10), which is equal to 85. In this way, although a higher KPI1 (production cost variation rate) can be obtained, it may cause a decrease in KPI2 (scarce equipment capacity), KPI3 (automated production line utilization rate), and / or KPI4 (balance among plants), resulting in a lower overall allocation score. For example, plant D may not be assigned to any production product. According to the greedy rule (2), the cost difference is the smallest among the cost differences, and the increase in cost difference will be the smallest when the production plant of product 1 is changed. Therefore, in the second operation of the greedy algorithm, the calculation unit 102 can select plant B to manufacture product 1, and keep the plants manufacturing products 2, 3, 4, and 5 as A, A, A, and B, respectively. In this way, the total production cost increases to (20+10+20+30+10), which is equal to 90. However, since the products are all assigned to plants A and B, and plants C and D are not assigned, the final overall allocation score may still be too low. Therefore, the calculation unit 102 can continue to execute the third operation of the greedy algorithm. Since the second-ranked product in the cost difference priority order could be product 2, product 3, or product 4, according to the greedy rule (1), the cost 20 of changing product 2 from factory D takes precedence over the cost 30 of changing product 3 from factory B, and also over the cost 40 of changing product 4 from factory C. Therefore, when performing the third operation of the greedy algorithm, the calculation unit 102 can choose factory D to manufacture product 2, and maintain the factories manufacturing products 1, 3, 4, and 5 as B, A, A, and B, respectively. The calculation unit 102 can continue to greedily search for the best local solution of KPI1 (production cost variation rate) to determine the global best solution (i.e., find the highest total allocation score).

[0167] In another embodiment, the computing unit 102 can execute metaheuristic algorithms, which may include genetic algorithms (GA), simplified swarm optimization (SSO), and tabu search algorithms, but the invention is not limited thereto. Genetic algorithms are algorithms that draw on phenomena in evolutionary biology (including heredity, mutation, selection, and mating) and can be applied to intelligent production planning. The computational units used by the computing unit 102 when executing genetic algorithms may include genes, chromosomes, and populations. For example, the computing unit 102 lists different products as different genes, different production machines (or plant areas) as chromosomes, and all products requiring matching in the same production plan as populations. The actions of the computing unit 102 when executing genetic algorithms may include "selection," "mating," and "mutation." "Selection" involves removing undesirable genes and retaining good ones based on the fitness values ​​corresponding to different chromosomes. "Mating" involves pairing genes to produce new chromosomes. "Mutation," on the other hand, involves altering certain genes within a selected chromosome to avoid the final result falling into a regionally optimal solution. Since not all products can be produced in all factories—meaning some products cannot be produced in connected factories and can only be produced in a single factory—optimal matching can only be performed on products with connected factories. After a preset number of iterations of the gene algorithm, the computation unit 102 can find combinations corresponding to higher total matching scores.

[0168] In another embodiment, the computing unit 102 can execute a simplified swarm optimization (SSO) algorithm, whose basic elements may include: a target, particles, particle best solutions (pbest), global best solutions (qbest), and updated solutions. When executing the SSO algorithm, the computing unit 102 predefines three parameters c. w c p c g This is used to specify how the solution should be updated next. For example, when ρ is between 0 and c... w If ρ is between c, the next updated solution will remain the current solution; when ρ is between c p to c g Between these, the next updated solution is adjusted to gbest, and so on. The solution is continuously adjusted by simulating the movement of each particle in the solution space and based on its individual pbest and global gbest.

[0169]

[0170] This is the current solution;

[0171] p ij The optimal solution for the particle (pbest);

[0172] g j The global optimal solution (qbest);

[0173] x is a random solution.

[0174] In another embodiment, the computing unit 102 can execute a tabu search algorithm, which is a method for finding approximate solutions to optimization problems using a local search method. The tabu search algorithm involves moving from the current solution to neighboring solutions, and its update mechanism uses a swap method to exchange the positions of two elements in the current solution. This "swap" operation refers to making a specific change to the current solution, typically swapping the positions of two elements, to generate a new solution.

[0175] In another embodiment, the computing unit 102 may perform linear programming, which may include multi-objective linear programming (MOLP), which finds the optimal solution by maximizing (or minimizing) multiple linear objective functions on a convex polyhedron. Assume the total allocation score is denoted by F and can be expressed by the following formula:

[0176] F = c1x1 + c2x2 + ... + c n x n

[0177] Where x1, x2, ..., x n These are the variables that need to be determined in order to achieve the purpose of the adaptive function (maximization or minimization). In this invention, x1, x2, ..., x... n These can be KPI1 (production cost), KPI2 (capacity of scarce equipment), KPI3 (automated line utilization rate), and KPI4 (balance between plants). b1, b2, ..., b m Then the following inequalities are satisfied:

[0178] a 11 x1+a 12 x2 + ... + a 1n x n ≤b1

[0179] a 21 x1+a 22 x2 + ... + a 2n x n ≤b2

[0180] a 31 x1+a32 x2 + ... + a 3n x n ≤b3

[0181] …

[0182] a m1 x1+a m2 x2 + ... + a mn x n ≤b m

[0183] a 11 To a mn b1 to b m c1 to c n These are all constants (e.g., other requirements and constraints such as capacity, demand, cost, profit, etc.).

[0184] Assuming the current goals are KPI1 (production cost) and KPI4 (factory balance), there are currently 6 products to be scheduled, and three factories can produce them. Table 34 below shows the constraints and production conditions for each product:

[0185] (Table 34)

[0186]

[0187] In the linear programming method, calculation unit 102 sets x1, x2, ..., x6 to represent six products respectively; and sets y1, y2, y3 to represent the demand allocation of the three plant areas respectively, and obtains the following inequalities:

[0188] y1≥60 (1)

[0189] 0≤y2≤80 (2)

[0190] 0≤y3≤240 (3)

[0191] y1+y2+y3≤240 (4)

[0192] Since the current objectives are KPI1 (production cost) and KPI4 (balance among factories), the calculation unit 102 solves the above inequalities (1) to (4) using linear programming and converges to the optimal solution. Among them, the factory area in which each product is placed is the arrangement result with the highest score in this round.

[0193] In another embodiment, computing unit 102 can perform deep learning, specifically a Q-learning algorithm utilizing a deep convolutional neural network. Q-learning is based on a Q-table and helps the agent learn better and achieve goals through self-learning interaction and Q-table updates. Q-learning can address the problem of excessively large Q-tables in general Q-learning when dealing with large or high-dimensional state spaces by using deep learning. Q-learning is a method of reinforcement learning whose purpose is to record learned policies to tell the agent which actions to take in different states to obtain the maximum reward, forming a Q-table. The agent uses the Q-table to interact with the environment.

[0194] Suppose there are currently 3 products to be produced, and there are 3 factories that can manufacture them. Table 35 below shows the limitations and production conditions for each product:

[0195] (Table 35)

[0196]

[0197] When performing Q-learning, the computing unit 102 can apply a0, a1, ..., a... to the information in Table 35. n The following actions yield Table 36:

[0198] (Table 36)

[0199]

[0200] In this context, action a0 indicates that calculation unit 102 selects factory A to produce product 1, factory B to produce product 2, and factory C to produce product 3. Action a1 indicates that calculation unit 102 selects factory B to produce product 1, factory B to produce product 2, and factory C to produce product 3. By executing the actions in Table 36, calculation unit 102 can obtain Table 37, which shows the states before and after each action:

[0201] (Table 37)

[0202] Action State after this action is performed State after the next action is performed [a0] [20,25,15,120,200,150] [15,45,0,105,360,0] [a1] [15,45,0,105,360,0] [20,0,40,120,0,280] … … … a n ]]> [0,45,15,0,360,150] [0,45,15,0,360,150]

[0203] Corresponding to Table 37, the calculation unit 102 can obtain the following Table 38 to represent the reward value corresponding to each action:

[0204] (Table 38)

[0205] Action Plant where production is performed Reward value a0 A, B, C -10,+6 a1 B, B, A +5,-2 … … … an B, B, C -2,+7

[0206] The calculation unit 102 of the capacity management system 100 achieves intelligent engine calculation through the above-mentioned "KPI logic adjustment", "KPI weight optimization" and "multi-objective optimization" to optimize the production quantity allocated to each plant area and maximize capacity.

[0207] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An energy management system, characterized by, The capacity management system comprises: a receiving unit configured to receive a plurality of production targets and a capacity supply amount of a first production machine and a capacity supply amount of a second production machine; a calculating unit configured to calculate a plurality of total allocation scores of the plurality of production targets allocated to the first production machine and the second production machine in different quantities; a comparing unit configured to compare the plurality of total allocation scores and select a maximum total allocation score; and an output unit configured to output quantities of the plurality of production targets allocated to the first production machine and the second production machine respectively corresponding to the maximum total allocation score.

2. The energy management system of claim 1, wherein, Each of the plurality of total allocation scores is calculated by multiplying a production cost variation rate by a first weight value and adding a device capacity load rate of the second production machine multiplied by a second weight value.

3. The energy management system of claim 2, wherein, The calculating unit calculates the first weight value and the second weight value by an objective weight method.

4. The energy management system of claim 3, wherein, The objective weight method comprises an entropy weight method or a principal component analysis method.

5. The energy management system of claim 2, wherein, The calculating unit calculates the first weight value and the second weight value by a subjective weight method.

6. The energy management system of claim 5, wherein, The subjective weight method comprises an expert experience method, an analytic hierarchy process or a decision experiment and evaluation laboratory method.

7. The energy management system of claim 1, wherein, The calculating unit further comprises calculating quantities of the plurality of production targets allocated to the first production machine and the second production machine by a greedy algorithm before calculating the plurality of total allocation scores.

8. The energy management system of claim 1, wherein, The calculating unit further comprises calculating quantities of the plurality of production targets allocated to the first production machine and the second production machine by a genetic algorithm before calculating the plurality of total allocation scores.

9. The energy management system of claim 1, wherein, The calculating unit further comprises calculating quantities of the plurality of production targets allocated to the first production machine and the second production machine by a simplified population algorithm before calculating the plurality of total allocation scores.

10. The energy management system of claim 1, wherein, The calculating unit further comprises calculating quantities of the plurality of production targets allocated to the first production machine and the second production machine by a tabu search algorithm before calculating the plurality of total allocation scores.

Images