Scheduling optimization method and apparatus using reinforcement learning

Abstract

This method for scheduling a naphtha decomposition center by at least one processor may comprise the steps of: acquiring input information; determining warehousing tank information by using a first agent on the basis of the input information; determining mixed tank combination information by using a second agent; and determining cracking furnace operation information by using a third agent.

Description

Method and device for scheduling optimization using reinforcement learning

[0001] The present disclosure relates to a method and apparatus for performing optimized scheduling using reinforcement learning.

[0002] Reinforcement learning is a branch of machine learning that describes how a defined agent within an environment recognizes its current state and selects the action or sequence of actions that maximizes reward from among available actions. Because reinforcement learning is a universally applicable solution, it is being studied in a variety of fields, and is particularly widely used in manufacturing system scheduling.

[0003] There has been research on optimization using mixed-integer linear programming (MILP), a type of mathematical optimization technique related to manufacturing system scheduling. However, due to the diverse and mathematically difficult to define production constraints in factory settings, this method has not been applied in practice. The present disclosure aims to address these issues.

[0004] The present disclosure provides a method and device for performing optimized scheduling through reinforcement learning using multiple agents.

[0005] One embodiment of the present disclosure can provide a method and device for performing optimized scheduling using reinforcement learning.

[0006] One embodiment of the present disclosure provides a method for scheduling a naphtha cracking center by at least one processor, comprising: obtaining input information; determining, based on the input information, receiving tank information using a first agent; determining mixing tank combination information using a second agent; and determining cracking operation information using a third agent, wherein at least one of the first agent, the second agent, and the third agent is learned by reinforcement learning.

[0007] In one embodiment, the first agent, the second agent, and the third agent may be asynchronous multi-agents.

[0008] In one embodiment, the input information may include at least one of constraints, naphtha receipt schedule information, tank inventory information, naphtha property information within the tank, mixing tank operation information, cracking furnace operation plan information, target production volume information for a specific product, raw material unit price information, and product unit price information.

[0009] In one embodiment, the method may include generating one or more scheduling information of the naphtha cracking center based on the receiving tank information, the mixing tank combination information, and the cracking furnace operation information.

[0010] In one embodiment, the input information may be obtained through a first UI (User Interface), and the scheduling information may be provided to the user through a second UI.

[0011] In one embodiment, the scheduling information may include at least one of: receiving scheduling information, mixing scheduling information, cracking scheduling information, expected production volume information, expected revenue information, expected naphtha inventory information, expected properties information, constraint satisfaction test result information, and a scheduling graph.

[0012] In one embodiment, the receiving tank information may include at least one of an identifier of at least one receiving tank for storing naphtha among a plurality of receiving tanks, naphtha receiving ratio information for each of the at least one receiving tank, and naphtha receiving schedule information for each of the at least one receiving tank.

[0013] In one embodiment, the mixing tank combination information may include a step of determining at least one of an identifier of at least one receiving tank from among a plurality of receiving tanks to which naphtha is to be transferred, information on a naphtha ratio to be transferred to the mixing tank for each identifier of the at least one receiving tank, information on a mixing schedule with the mixing tank for each of the at least one receiving tank, information on a naphtha mixing ratio for each of the at least one receiving tank, and information on a blending performance date.

[0014] In one embodiment, the decomposition furnace operation information may include at least one of decomposition furnace mode information, decomposition furnace identifier, input speed, Coil Outlet Temperature (COT), Dilution Steam Ratio (DSR), heating time, and decomposition furnace operation schedule information.

[0015] In one embodiment, at least one of the first agent, the second agent, and the third agent may be plural.

[0016] In one embodiment, the first agent, the second agent, and the third agent can be trained using the same reward during reinforcement learning.

[0017] In one embodiment, the reward of the reinforcement learning may be determined based on total revenue, facility operating costs, naphtha purchase costs, and costs according to constraints.

[0018] In one embodiment, the reward of the reinforcement learning is:

[0019]

[0020] , where profit is determined based on total revenue minus facility operating costs and naphtha purchase costs, w c is the weight per constraint, and Cost c may be the cost per constraint.

[0021] One embodiment of the present disclosure comprises a system comprising one or more processors; and one or more memories having stored thereon computer-readable instructions configured to cause the one or more processors to perform a method according to one embodiment of the present disclosure.

[0022] One embodiment of the present disclosure includes a program stored on a recording medium to cause a computer to execute a method according to one embodiment of the present disclosure.

[0023] One embodiment of the present disclosure includes a computer-readable recording medium having recorded thereon a program for executing a method according to one embodiment of the present disclosure on a computer.

[0024] One embodiment of the present disclosure includes a computer-readable recording medium having recorded thereon a database used in one embodiment of the present disclosure.

[0025] According to the present disclosure, optimized scheduling may be possible in various facilities such as naphtha cracking centers.

[0026] FIG. 1 is a schematic diagram of a product production process of a naphtha cracking center according to one embodiment of the present disclosure.

[0027] FIG. 2 is a diagram illustrating a reinforcement learning method using multiple agents according to one embodiment of the present disclosure.

[0028] FIG. 3a and FIG. 3b are diagrams illustrating a reinforcement learning method using multiple agents according to one embodiment of the present disclosure.

[0029] FIG. 4 is a flowchart illustrating a method for generating scheduling information according to one embodiment of the present disclosure.

[0030] FIG. 5 is a drawing showing an effect according to one embodiment of the present disclosure.

[0031] FIG. 6 is a block diagram of a system according to one embodiment of the present disclosure.

[0032] To clarify the technical idea of ​​the present disclosure, embodiments of the present disclosure will be described in detail with reference to the attached drawings. In describing the present disclosure, if a detailed description of a related known function or component is determined to unnecessarily obscure the gist of the present disclosure, the detailed description will be omitted. Components having substantially the same functional configuration among the drawings are given the same reference numbers and symbols as possible even if they are shown in different drawings. For convenience of explanation, devices and methods are described together when necessary. Each operation of the present disclosure does not necessarily have to be performed in the described order and may be performed in parallel, selectively, or individually.

[0033] The terms used in the embodiments of this disclosure have been selected from widely used, current terms, taking into account the functions of the present disclosure. However, these terms may vary depending on the intentions of those skilled in the art, precedents, the emergence of new technologies, etc. Furthermore, in certain cases, the applicant may arbitrarily select terms, and in such cases, their meanings will be described in detail in the description of the relevant embodiments. Therefore, the terms used in this specification should not be defined simply as names of terms, but rather based on their meanings and the overall content of the present disclosure.

[0034] Throughout this disclosure, singular expressions may include plural expressions unless the context clearly dictates otherwise. Terms such as "comprise" or "have" should be understood to indicate the presence of a feature, number, step, operation, component, part, or combination thereof, but do not preclude the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof. In other words, when it is said throughout this disclosure that a part "comprises" a certain component, unless specifically stated otherwise, this does not mean that other components may be included, but rather that other components may be excluded.

[0035] Expressions such as "at least one" modify the entire list of elements, not individual elements of the list. For example, "at least one of A, B, and C" and "at least one of A, B, or C" refer to only A, only B, only C, both A and B, both B and C, both A and C, all of A, B, and C, or any combination thereof.

[0036] In addition, terms such as “...unit”, “...module”, etc. described in the present disclosure mean a unit that processes at least one function or operation, which may be implemented as hardware or software, or a combination of hardware and software.

[0037] Throughout this disclosure, when a part is said to be "connected" to another part, this includes not only cases where the parts are "directly connected," but also cases where the parts are "electrically connected" with other elements intervening. Furthermore, when a part is said to "include" a component, this does not exclude other components, but rather includes other components, unless otherwise specifically stated.

[0038] The expression “configured to” as used throughout this disclosure can be used interchangeably with, for example, “suitable for,” “having the capacity to,” “designed to,” “adapted to,” “made to,” or “capable of.” The term “configured to” does not necessarily mean something that is “specifically designed to” in terms of hardware. Instead, in some contexts, the expression “a system configured to” can mean that the system, together with other devices or components, is “capable of.” For example, the phrase “a processor configured (or set) to perform A, B, and C” may mean a dedicated processor (e.g., an embedded processor) for performing those operations, or a generic-purpose processor (e.g., a CPU or application processor) that can perform those operations by executing one or more software programs stored in memory.

[0039] The present disclosure provides a method for performing optimized scheduling using reinforcement learning. For convenience, the method for performing scheduling at a naphtha cracking center will be described as an example. However, the embodiments of the present disclosure are not limited to naphtha cracking centers and can be applied to optimized scheduling of other facilities as well. Throughout this disclosure, naphtha may also be referred to as naphtha.

[0040] At the Naphtha Cracking Center (NCC), naphtha, a gasoline fraction obtained from an atmospheric distillation unit of crude oil, is thermally cracked in a high-temperature cracking furnace, and then through processes such as rapid cooling, compression, and refining, ethylene, propylene, butylene, and BTX (Benzene, Toluene, and Xylene), which are basic raw materials for petrochemical products, can be produced.

[0041] That is, naphtha is converted into substances with high industrial utility, such as ethylene, propylene, benzene, toluene, and xylene, through steam or thermal cracking. For example, ethylene becomes a raw material for polyethylene and polystyrene, propylene becomes polypropylene, and butane or butylene can be used to make synthetic rubber. These substances serve as raw materials for the plastics processing, textile, rubber, paint, and detergent industries, and these raw materials can be transformed into final products, such as daily necessities, adhesives, dyes, pesticides, pharmaceuticals, industrial products, and interior design materials.

[0042] FIG. 1 is a schematic diagram of a product production process of a naphtha cracking center according to one embodiment of the present disclosure.

[0043] Referring to Figure 1, a naphtha cracking center is a core facility that produces petrochemical raw materials through a complex process. It consists of a receiving stage for unloading naphtha, a mixing stage for mixing naphtha, and a cracking stage for producing marketable products. More specifically, the naphtha cracking center initially transports naphtha from geographically distributed refineries via ships and other means, unloads it into receiving tanks, and then supplies various naphthas from various receiving tanks to a mixing tank. The mixed naphtha in the mixing tank is then heated in a cracking furnace to produce marketable products of the desired quality. That is, the product production process of the naphtha cracking center may include a receiving process of storing naphtha supplied from one or more ships (110) or companies (e.g., other oil companies) in one or more receiving tanks (120), a mixing process of transferring the naphtha in the receiving tanks (120) to a mixing tank (130) for a naphtha cracking process, and a cracking process of subjecting the naphtha supplied from the mixing tank (130) to high-temperature thermal cracking in a cracking furnace (140). Here, the mixing tank (130) may also be referred to as a blending tank or a feed tank.

[0044] In one embodiment, the product production process of the naphtha cracking center may further include a process of measuring the paraffin content of naphtha supplied from a vessel (110) or a company, and a process of measuring the paraffin content of naphtha stored in a receiving tank (120), a mixing tank (120), etc.

[0045] In one embodiment, the constraints may include a range for the paraffin content of each tank. For example, the paraffin content of the mixing tank (130) may need to be limited to a range of approximately 80 to 83% based on the total weight of the naphtha. Since naphtha has different properties depending on the origin or company, the naphtha stored in the receiving tank also has different properties, and the receiving and mixing processes must be performed so that the paraffin content of the mixing tank (130) satisfies the constraint range.

[0046] Considering these constraints, developing an optimal schedule for naphtha cracking centers is crucial for profitability and efficiency. Currently, experts rely on their experience and expertise to determine which receiving tanks to store naphtha in, the ratio at which naphtha will be blended with the mixing tank, and the cracking furnace to heat to a certain degree. This approach has limitations in predicting complex chemical reactions and actual outcomes, and can vary significantly depending on the expert's experience and expertise. It's also difficult to verify that all constraints have been met, and it's difficult to respond to sudden changes in circumstances.

[0047] Accordingly, the present disclosure aims to provide a method for determining optimal scheduling using artificial intelligence. Specifically, the present disclosure aims to provide a method for autonomously operating a naphtha cracking center using multi-agent reinforcement learning, where each agent takes responsibility for each step and cooperates to achieve a common goal, overcoming real-world constraints. The artificial intelligence-related functions according to the present disclosure are operated via a processor and memory. The processor may be comprised of one or more processors. In this case, one or more processors may be general-purpose processors such as a CPU, AP, or DSP (Digital Signal Processor), graphics-dedicated processors such as a GPU or VPU (Vision Processing Unit), or AI-dedicated processors such as an NPU. The one or more processors control the processing of input data according to predefined operating rules or AI models stored in memory. Alternatively, if one or more processors are AI-dedicated processors, the AI-dedicated processors may be designed with a hardware structure specialized for processing a specific AI model.

[0048] FIG. 2 is a diagram illustrating a reinforcement learning method using multiple agents according to one embodiment of the present disclosure.

[0049] Referring to FIG. 2, a naphtha cracking center can be operated according to an optimal schedule using scheduling information determined through reinforcement learning using multiple agents (210, 220, 230). A simulator using reinforcement learning can take actions (240, 250, 260) from agents (210, 220, 230) and provide the next observation and reward (280) based on the current action. In one embodiment, each agent is responsible for a specific process and can cooperate with each other to achieve a goal such as profit maximization while complying with realistic constraints. For example, realistic constraints may exist such as performing the transfer process from the receiving tank to the mixing tank for at least 8 hours while not exceeding the minimum naphtha storage capacity of the receiving tank and the maximum naphtha storage capacity of the mixing tank.

[0050] In one embodiment, each agent can determine the information necessary to generate scheduling information for a naphtha cracking center for a predetermined period of time in the future based on current information and various constraints, such as the inventory status of each tank, vessel arrival plans, naphtha supply plans from other companies, and prices of naphtha and marketable products.

[0051] In one embodiment, each of the multiple agents can produce different results (e.g., durations) at different times. For example, the first agent (210) is an agent that manages receiving, and when ships arrive irregularly, determines actions such as selecting a receiving tank to store naphtha and determining the amount to store in that receiving tank. The second agent (220) is an agent that blends naphtha, and when the level of a receiving tank reaches a threshold (e.g., 90% of the tank capacity), determines actions such as selecting a receiving tank to bring naphtha into the blending tank and determining the amount to bring from that receiving tank. In addition, the third agent (230) is an agent that manages a cracking furnace, and when the product inventory falls below a predetermined amount, determines actions such as determining variables for receiving naphtha from the blending tank and operating the cracking furnace. A virtual NAFTA operating environment (270) can be created with actions (240, 250, 260) determined at different times. The simulation device can determine the expected revenue in the virtual NAFTA operating environment (270) and determine a reward (280) based on the expected revenue. This reward can be delivered to multiple agents (210, 220, 230) and used by the agents to perform reinforcement learning. That is, multiple agents (210, 220, 230) can learn using the same reward during reinforcement learning. However, it is also possible for multiple agents to learn using different rewards.

[0052] In one embodiment, the reward (280) may be determined based on total revenue, facility operating costs, naphtha purchase costs, and costs subject to constraints. For example, the reward may be determined by [Mathematical Formula 1] below.

[0053] [Mathematical Formula 1]

[0054]

[0055] In [Mathematical Formula 1], Constraints are constraints, and w c is the weight per constraint, and Cost c can mean the cost per constraint. Accordingly, the more constraints are violated, the higher the cost. c The value can be increased. For example, if the constraint includes the stability of the paraffin component, i.e., the condition that the composition of the paraffin must be maintained to a certain degree, the change in the composition of the paraffin stored in the mixing tank can be used as the variable c.

[0056] Additionally, in [Equation 1], profit can be calculated by subtracting the expected production cost of marketable products from the expected revenue generated from selling naphtha, taking into account facility operating costs (e.g., energy usage costs) and naphtha purchase costs. For example, profit can be determined according to [Equation 2], which subtracts facility operating costs and naphtha purchase costs from total revenue, as shown below.

[0057] [Equation 2]

[0058]

[0059] In one embodiment, Revenue may be calculated as, for example, "CH4 Production * CH4 Product Price + PSA OFF GAS Production * PSA OFF GAS Price + RC2 Production * Ethane Product Price + C3LPG Production * Propane Product Price + Ethylene Production * C2 Product Price + Propylene Production * C3 Product Price + H2(99%) Production * 99% H2 Product Price + HRPG Production * HRPG Product Price + PFO Production * PFO Product Price + RawC5 Production * RawC5 Product Price + (Mixed C4 Production * Mixed C4 Product Price) + (RPG Production RawC5 Production) * RPG Product Price."

[0060] In one embodiment, Energy usage can be calculated as, for example, "[(Naphtha input + C3LPG input + C4LPG input) * A + Mixed C4 production * B + (RPG production RawC5 production) * C ] * C3LPG price / C3LPG calorific value / 1000", where A can be the average energy intensity of a naphtha cracking center plant, B can be the average energy intensity of a BD plant, and C can be the average energy intensity of a BTX plant (e.g., a plant that produces aromatic products using pyrolysis gasoline produced in an ethylene plant as a raw material).

[0061] In one embodiment, hh may be calculated as "Total naphtha feed input * naphtha price + C3 LPG input * C3 LPG price + C4 LPG input * C4 LPG price + RC2 input * ethane product price".

[0062] In one embodiment, once an optimal schedule is determined through reinforcement learning, the naphtha cracking center can be operated based on the generated optimal schedule. For example, if multiple schedules are created and provided to a user, the user can operate the naphtha cracking center based on one of them.

[0063] FIG. 3a and FIG. 3b are diagrams illustrating a reinforcement learning method using multiple agents according to one embodiment of the present disclosure.

[0064] Referring to Figure 3a, an asynchronous multi-agent system is illustrated in which the action start and duration times of each agent are different. For example, if the multi-agents are composed of three agents, a first agent (310), a second agent (320), and a third agent (330), as shown in Figure 3a, each agent decides on a different action at a different time, and the decided action vector The state vector is changed as a result of the actions being transmitted to the environment (340) and applying the actions to the environment (340). can be determined. Also, the action termination vector can be provided to agents. In addition, rewards generated as a result of applying actions to the environment (340) Each agent can also be provided with a task. The actions of each agent can be determined asynchronously, as shown in Fig. 3b.

[0065] In another embodiment, the actions of each agent may be transmitted to the environment (340) each time the action is determined. For example, at the first time when the first agent (310) determines the first action, the first action is reflected in the environment (340), at the second time when the second agent (310) determines the second action, the second action is reflected in the environment (340), and at the third time when the second agent (310) determines the third action, the third action is reflected in the environment (340), and so on. However, since the reward is determined only when the product is ultimately produced and sold, the reward may be determined and provided to each agent after the actions of the first agent (310), the second agent (320), and the third agent (330) are all reflected in the environment (340).

[0066] In one embodiment, the state, actions, and rewards of each agent may include the following information:

[0067] 1. First Agent

[0068] - Status: naphtha receiving schedule (e.g. vessel schedule or receiving plan from another company), current status of receiving tanks (e.g. naphtha inventory and properties per tank), constraints associated with receiving tanks.

[0069] - Action: Identifier of the receiving tank to store naphtha when receiving naphtha, the amount of naphtha to be stored in the receiving tank, or the pipe connection schedule (e.g., pipe connection to vessel A from 2:00 PM to 10:00 PM for receiving tank No. 1).

[0070] - Reward: Profit considering whether constraints are satisfied

[0071] 2. Second Agent

[0072] - Status: Stock and properties of each receiving tank, stock and properties of the mixing tank

[0073] - Action: Identifier of the receiving tank that will hold at least some naphtha in the mixing tank, the amount of naphtha to be brought from that receiving tank to the mixing tank, or the pipe connection schedule (e.g., pipe connection from receiving tank 1 to the mixing tank from 8:00 a.m. to 2:00 p.m.).

[0074] Compensation: Profits that take into account whether constraints are satisfied.

[0075] 3. Third Agent

[0076] Status: Inventory and properties of mixing tanks, operation status by each disassembly

[0077] Action: Feeding speed, COT, DS ratio, etc. of naphtha from the mixing tank to each cracker

[0078] Compensation: Profits that take into account whether constraints are satisfied.

[0079] However, this is just an example, and it is obvious that the state, actions, and rewards of each agent can be adjusted differently.

[0080] FIG. 4 is a flowchart illustrating a method for generating scheduling information according to one embodiment of the present disclosure.

[0081] Referring to FIG. 4, in operation 410, the naphtha cracking center scheduling system may obtain input information. For example, the input information may be obtained through user input based on a user interface (UI).

[0082] In one embodiment, the input information may include constraints, naphtha receipt schedule information, tank inventory information, naphtha property information within the tank, mixing tank operation information, cracking furnace operation plan information, target production volume information for a specific product, raw material unit price information, product unit price information, etc. In another embodiment, some of the information, such as constraints, may be preset, and in that case, may not be included in the input information because it has been preset.

[0083] In one embodiment, constraints may include physical constraints such as tank storage capacity criteria that must be satisfied, the number of pipes that can be connected at one time, etc., stability constraints regarding stability, operational constraints to meet a set target production amount for a specific period of time (e.g., weekly or monthly), etc. In addition, naphtha receipt schedule information may include a vessel receipt schedule, tank information of other companies, naphtha receipt schedule for a specific period in the future, expected receipt date and time, receipt speed, receipt amount, naphtha property information, identification information according to naphtha receipt method (e.g., vessel identifier, tank identifier of other companies, pipe identifier by company, etc.).

[0084] In one embodiment, at least one of the input information may include information for a specific period of time or information at a specific point in time. For example, the tank inventory information and the naphtha properties information within the tank may each include naphtha inventory information for the corresponding tank at the start of the scheduling and naphtha properties information for the corresponding tank at the start of the scheduling.

[0085] In one embodiment, the blending tank operation information may include one or more blending schedules including a blending start time, a blending end time, the names of the receiving tanks to be blended, and a blending speed for each blending receiving tank. For example, the blending schedule may include a recent blending schedule.

[0086] In one embodiment, the furnace operation plan information may include schedule information for each furnace for a predetermined period in the future. For example, the furnace operation plan information may include schedule information determined for each furnace for the next 30 days. In one embodiment, the furnaces may exist in various types. For example, if there are a first furnace, a second furnace, and a third furnace of different types, the furnace operation plan information may include schedule information determined for each of the first furnace, the second furnace, and the third furnace for the next 30 days. The furnace operation plan information may include a furnace start time, a furnace end time, an operation mode (or feed mode), decoking schedule information, COT (coil outlet temperature), coil outlet pressure, a predetermined speed (e.g., feed rate), DS (Dilution Steam) ratio, etc.

[0087] In one embodiment, target production volume information for a specific product may include a target production volume or rate for a specific product over a specific period of time. For example, target production volume information for a specific product may include a daily ethylene production target for the next 30 days, a daily propylene production target for the next 30 days, etc.

[0088] In one embodiment, raw material unit price information may include raw material unit price information at the time of information input, raw material unit price information for a specific period prior to the time of information input, and projected raw material unit price information for a specific period after the time of information input. For example, raw material unit price information may include the projected daily price of raw materials for the next 30 days.

[0089] In one embodiment, the product unit price information may include unit price information for each product at the time of entry, unit price information for a specific period prior to the time of entry, and projected unit price information for a specific period after the time of entry. For example, the product unit price information may include the projected daily price of naphtha products for the next 30 days.

[0090] However, the above input information is only an example and is not limited thereto, and various input information for scheduling the naphtha cracking center may be included.

[0091] In operation 430, the naphtha cracking center scheduling system may determine storage tank information using a first agent based on input information. In one embodiment, the first agent may be an agent trained using reinforcement learning. The first agent may determine a storage tank to store naphtha from at least one of a vessel and a tank of another company based on input information including a vessel storage schedule, a naphtha storage schedule from another company, real-time storage tank inventory, information on the properties of naphtha in a tank, and information on a cracking furnace operation plan, and may determine information on the amount, rate, or schedule of naphtha to be stored in the corresponding storage tank. For example, the first agent may determine an identifier of at least one storage tank among a plurality of storage tanks to store naphtha based on the input information, and determine information on the naphtha storage rate or amount for each tank corresponding to each identifier. In addition, the first agent may determine, based on the input information, naphtha storage schedule information for each tank corresponding to each identifier, information on the storage period of naphtha in the corresponding storage tank, and the like. The receiving schedule information may include date and time information for connecting a vessel or other company's equipment and a receiving tank via a pipe (e.g., naphtha receiving from vessel B to receiving tank A from 2:00 PM to 6:00 PM). In one embodiment, the pipe may be connected by transferring naphtha from the vessel or other company's equipment to the receiving tank. However, since pipe connection and the like are inconvenient for humans to perform if the schedule changes frequently, constraints such as a minimum n-hour connection per pipe may exist. The first agent may determine the receiving tank information by considering these constraints.

[0092] In one embodiment, the naphtha cracking center scheduling system can obtain naphtha properties information corresponding to each receiving tank after a predetermined amount of naphtha has been distributed to the receiving tank.

[0093] In one embodiment, all receiving tank information may be determined using a single first agent, receiving tank information may be determined using a different first agent for each receiving tank, or receiving tank information may be determined using a different first agent for each group of receiving tanks. That is, there may be more than one first agent.

[0094] In operation 450, the naphtha cracking center scheduling system may determine mixing tank combination information using a second agent. In one embodiment, the second agent may be an agent trained using reinforcement learning. The second agent may determine mixing tank combination information based on the inventory of each receiving tank, the properties of the naphtha stored in each receiving tank, and other factors.

[0095] In one embodiment, the blending tank combination information may include an identifier of at least one receiving tank from among a plurality of receiving tanks to which naphtha is to be transferred, information on a ratio (or quantity) of naphtha to be transferred to the blending tank for each identifier of the at least one receiving tank, information on a blending schedule with the blending tank for each of the at least one receiving tank, information on a naphtha blending ratio for each of the at least one receiving tank, information on a blending execution date, etc. The naphtha blending ratio information may include information on a ratio or quantity of naphtha to be taken from each receiving tank, etc.

[0096] In one embodiment, all mixed tank combination information may be determined using a single second agent, the receiving tank information may be determined using a different second agent for each mixed tank, or the receiving tank information may be determined using a different second agent for each group of mixed tanks. That is, there may be more than one second agent.

[0097] In operation 470, the naphtha cracking center scheduling system may determine cracking operation information using a third agent. In one embodiment, the third agent may be an agent trained using reinforcement learning. The third agent may determine cracking operation information based on inventory information of a mixing tank, property information of the mixing tank, cracking furnace status information, etc. In one embodiment, the cracking furnace operation information may include cracking furnace mode information, a cracking furnace identifier, a feed rate, a Coil Outlet Temperature (COT), a Dilution Steam Ratio (DSR), a heating time, cracking furnace operation schedule information, one or more variables for cracking furnace operation, etc.

[0098] In one embodiment, there may be multiple third agents. For example, different agents may be used for each decomposition mode. For example, the third agents may include a 3-1 agent that has learned reinforcement learning for the decomposition mode A, a 3-2 agent that has learned reinforcement learning for the decomposition mode B, and a 3-3 agent that has learned reinforcement learning for the decomposition mode C, and the naphtha cracking center scheduling system may determine the operation information of the decomposition mode A using the 3-1 agent for the decomposition mode A, determine the operation information of the decomposition mode B using the 3-2 agent for the decomposition mode B, and determine the operation information of the decomposition mode C using the 3-3 agent for the decomposition mode C. In another embodiment, one third agent may determine the operation information of all decomposition modes.

[0099] In one embodiment, the naphtha cracking center scheduling system may determine one or more scheduling information of the naphtha cracking center based on the receiving tank information generated by the first agent, the mixing tank combination information generated by the second agent, and the cracking furnace operation information generated by the third agent. In one embodiment, the scheduling information may include receiving scheduling information, mixing scheduling information, cracking furnace scheduling information, expected production volume information, expected profit information, expected naphtha inventory information, expected properties information, constraint satisfaction test result information, a scheduling graph, and the like.

[0100] In one embodiment, the receiving scheduling information may include information on receiving schedules for a predetermined period in the future. For example, the receiving scheduling information may include information on the identification of tanks to be received for the next two weeks, the start time of receiving for the corresponding receiving tank, and the end time of receiving for the corresponding receiving tank.

[0101] In one embodiment, the mixing scheduling information may include mixing schedule information for a predetermined period in the future. For example, the mixing scheduling information may include information on the identification of a mixing tank to be used for the next two weeks, the mixing start time of the mixing tank, the mixing end time of the mixing tank, and mixing speed information of the tank (e.g., the mixing speed of Tank A: approximately 100 tons / hour).

[0102] In one embodiment, the decomposition scheduling information may include decomposition schedule information for a predetermined period of time in the future. For example, the decomposition scheduling information may include the identification of the decomposition furnace to be used for the next two weeks, the decomposition start time for the decomposition furnace, the decomposition end time for the decomposition furnace, the decomposition speed (e.g., a target control speed determined through artificial intelligence, such as a feed rate), the cost-of-transaction (COT), and the die-saving ratio.

[0103] In one embodiment, projected production volume information, projected revenue information, projected naphtha inventory information, and projected properties information may also be projected information for a predetermined period in the future. For example, projected production volume information may include projected daily production by product for the next two weeks. Furthermore, projected naphtha inventory information may include information on naphtha inventory volume or naphtha change by tank for the next two weeks, and projected properties information may include information on property change by tank for the next two weeks.

[0104] In one embodiment, the constraint satisfaction test result information may include evaluation information on how well the generated schedule satisfies the predetermined constraints.

[0105] Furthermore, the naphtha cracking center scheduling system can provide users with one or more scheduling information via the UI / UX. For example, the UI / UX can provide an overview of each of the one or more scheduling information items in the form of a graph or diagram, and can also provide summary information such as cumulative profit and constraint satisfaction.

[0106] According to one embodiment of the present disclosure, a naphtha cracking center scheduling system may determine scheduling information for a naphtha cracking center using an asynchronous multi-agent system including a first agent, a second agent, and a third agent. For example, each agent may determine different information at different times.

[0107] Meanwhile, although FIG. 4 illustrates a schedule generation method according to one embodiment of the present disclosure, it is to be understood that various modifications may be made to FIG. 4. For example, although FIG. 4 illustrates sequential operations, it is to be understood that the various operations in FIG. 24 may overlap, be performed in parallel, be performed in a different order, or some operations may be performed repeatedly multiple times.

[0108] FIG. 5 is a drawing showing an effect according to one embodiment of the present disclosure.

[0109] Referring to Figure 5, it can be seen that when artificial intelligence using reinforcement learning according to one embodiment of the present disclosure schedules a naphtha cracking facility, profits increase by approximately 9.46% compared to when a field expert makes the decision. According to one embodiment of the present disclosure, profits can be maximized by having artificial intelligence generate an optimal schedule.

[0110] FIG. 6 is a block diagram of a system according to one embodiment of the present disclosure.

[0111] Referring to FIG. 6, a naphtha cracking center scheduling device (600) (the device may also be referred to as a server or a system) may include a transceiver (610), a memory (620), a database (630), and a processor (640). However, not all of the components illustrated in FIG. 6 are essential components of the naphtha cracking center scheduling device (600). The naphtha cracking center scheduling device (600) may be implemented with more components than the components illustrated in FIG. 6, or may be implemented with fewer components than the components illustrated in FIG. 6. In addition, the transceiver (610), the memory (620), and the processor (640) may be implemented in the form of a single chip.

[0112] In one embodiment, the transceiver (610) may communicate with a terminal or other electronic device connected to the naphtha cracking center scheduling device (600) via wired or wireless means. For example, the transceiver (610) may receive input information from a user terminal. In one embodiment, the input information may include naphtha stocking plan information, production target amount for each product (e.g., ethylene production target amount), constraints, status information, scheduling start time, etc. The naphtha stocking plan information may include expected stocking time, stocking speed, stocking amount, naphtha properties information, and stocking type information (e.g., whether it is a vessel or a tank of another company). The status information may include a preset mixing schedule, a preset cracking furnace schedule, etc. The mixing schedule or cracking furnace schedule may include a start time, an end time, a tank name, and a mixing or cracking speed for each tank, and the cracking furnace schedule may further include various variable information such as pressure and temperature information. In addition, the status information may include naphtha inventory amount and properties information for each tank at the time of scheduling start.

[0113] Various types of data, such as programs and files, such as applications, can be installed and stored in the memory (620). The processor (640) can access and use data stored in the memory (620), or store new data in the memory (620). In addition, the memory (620) can store one or more instructions. The processor (640) can execute one or more instructions stored in the memory.

[0114] The processor (640) controls the overall operation of the naphtha cracking center scheduling device (600) and may include at least one processor, such as a CPU or a GPU. The processor (640) may control other components included in the naphtha cracking center scheduling device (600) to perform operations for operating the naphtha cracking center scheduling device (600). For example, the processor (640) may determine storage tank information using a first agent, determine mixing tank combination information using a second agent, and determine cracking furnace operation information using a third agent based on input information.

[0115] In one embodiment, the processor (640) may generate one or more scheduling information of the naphtha cracking center based on the receiving tank information, the mixing tank combination information, and the cracking furnace operation information. In addition, the transceiver (610) may transmit the scheduling information to the user terminal so that the corresponding scheduling information is output through the display of the user terminal. In one embodiment, the output scheduling information may include a receiving schedule for a predetermined period in the future (e.g., receiving start and end times, receiving tank identifiers, etc.), a mixing schedule for a predetermined period in the future (e.g., mixing start and end times, mixing tank identifiers, mixing speed, etc.), a schedule for each cracking furnace for a predetermined period in the future (e.g., cracking start and end times, feed rate determined through an algorithm, COT, DS ratio, etc.), information on daily production volume and expected profit for each product for a predetermined period in the future, information on naphtha inventory and property change information for a predetermined period in the future, information on constraint inspection results for the generated schedule, a plot for visualizing the generated schedule, etc.

[0116] This process of generating and outputting scheduling information can be performed through the UI / UX of the user terminal. For example, when the processor (640) obtains input information input by the user, it verifies whether the input information contains sufficient data to generate output data. If the input information is determined to be valid, it can generate one or more scheduling information using an artificial intelligence scheduler based on the scheduling start date input by the user. In addition, the processor (640) can provide information by graphing one or more scheduling information, and the user terminal can display this information in the form of a UI / UX.

[0117] The database (630) can store various learning data for training the learning model. Furthermore, the database (630) may store material information, phase information, simulation result information, and the like. In various embodiments, the database may also store output data generated by the learning model. While FIG. 6 illustrates the naphtha cracking center scheduling device (600) as including the database (630), the database (630) may also be provided externally to the device. In this case, the database (630) may be connected to the naphtha cracking center scheduling device (600) via wired or wireless connections.

[0118] Additionally, the learning model may be implemented outside the naphtha cracking center scheduling device (600) (e.g., cloud-based) or may be included within the naphtha cracking center scheduling device (600).

[0119] An embodiment of the present disclosure may also be implemented in the form of a recording medium containing computer-executable instructions, such as program modules, executed by a computer. Computer-readable media may be any available media that can be accessed by a computer, and includes both volatile and nonvolatile media, removable and non-removable media. Furthermore, computer-readable media may include both computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Communication media typically contains computer-readable instructions, data structures, or program modules, and includes any information delivery media.

[0120] The above description of the present disclosure is provided for illustrative purposes only, and those skilled in the art will readily appreciate that the present disclosure can be readily modified into other specific forms without altering the technical spirit or essential characteristics of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, components described as being single may be implemented in a distributed manner, and similarly, components described as being distributed may be implemented in a combined manner.

[0121] The scope of the present disclosure is indicated by the claims described below rather than the detailed description above, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present disclosure.

Claims

1. A method for scheduling a naphtha cracking center by at least one processor, Step of obtaining input information; A step of determining the incoming tank information using a first agent based on the above input information; A step of determining mixed tank combination information using a second agent; and A step of determining operation information by decomposition using a third agent is included. At least one of the first agent, the second agent, and the third agent, A method characterized by learning by reinforcement learning.

2. In paragraph 1, the first agent, the second agent, and the third agent, A method characterized by being an asynchronous multi-agent.

3. In paragraph 1, the input information is: A method comprising at least one of constraints, naphtha receipt schedule information, tank inventory information, naphtha property information in the tank, mixing tank operation information, cracking furnace operation plan information, target production volume information for a specific product, raw material unit price information, and product unit price information.

4. In paragraph 1, the method, A method comprising the step of generating one or more scheduling information of the naphtha cracking center based on the receiving tank information, the mixing tank combination information, and the cracking furnace operation information.

5. In paragraph 4, the input information is: Obtained through the first UI (User Interface), The above scheduling information is, A method provided to a user through a second UI.

6. In paragraph 4, the scheduling information is: A method comprising at least one of receiving scheduling information, blending scheduling information, cracking scheduling information, expected production volume information, expected revenue information, expected naphtha inventory information, expected properties information, constraint satisfaction test result information, and a scheduling graph.

7. In paragraph 1, the storage tank information is: A method comprising at least one of an identifier of at least one receiving tank for storing naphtha among a plurality of receiving tanks, naphtha receiving ratio information for each of the at least one receiving tank, and naphtha receiving schedule information for each of the at least one receiving tank.

8. In paragraph 1, the mixing tank combination information is, A method comprising the step of determining at least one of an identifier of at least one receiving tank from among a plurality of receiving tanks to which naphtha is to be transferred, information on a ratio of naphtha to be transferred to the mixing tank for each identifier of the at least one receiving tank, information on a mixing schedule with the mixing tank for each of the at least one receiving tank, information on a naphtha mixing ratio for each of the at least one receiving tank, and information on a blending performance date.

9. In paragraph 1, the operation information of the decomposition furnace is, A method comprising at least one of decomposition furnace mode information, decomposition furnace identifier, feed rate, COT (Coil Outlet Temperature), DSR (Dilution Steam Ratio), heating time, and decomposition furnace operation schedule information.

10. In the first paragraph, at least one of the first agent, the second agent, and the third agent, A method characterized by a plurality of 11. In the first paragraph, the first agent, the second agent, and the third agent, A method characterized in that learning is performed using the same reward during reinforcement learning.

12. In the 11th paragraph, the reward of the reinforcement learning is A method determined based on total revenue, facility operating costs, naphtha purchase costs, and constraints.

13. In paragraph 11, the reward of the reinforcement learning is is determined by, Here, profit is determined based on total revenue minus facility operating costs and naphtha purchase costs. w c is the weight per constraint, and Cost c is the cost per constraint, method.

14. As a system, one or more processors; and A system comprising one or more memories having computer-readable instructions stored thereon configured to cause one or more processors to perform the method of any one of claims 1 to 13.

15. A program stored on a computer-readable recording medium that causes a computer to execute any one of the methods of clauses 1 to 13.

Images