Kitchen electricity group cooperative work intelligent scheduling control method and kitchen electricity group cooperative work intelligent scheduling control system

Abstract

The invention provides a kitchen appliance group cooperative work intelligent scheduling control method and system, and relates to the technical field of smart home, and the method comprises the steps: obtaining the operation state and cooking task information of kitchen appliance equipment, carrying out the process disassembly through combining with a cooking technology constraint rule, and forming a task execution graph; calculating a criticality quantitative index, and carrying out equipment distribution based on the feasibility mapping relation and the criticality quantitative index; the method identifies the combination of processes which can be executed in parallel, and distributes the equipment again to generate the scheduling sequence, thereby achieving the efficient cooperative work of kitchen electric equipment, and improving the execution efficiency of cooking tasks.

Description

Technical Field

[0001] This invention relates to the field of smart home technology, and in particular to an intelligent scheduling and control method and system for collaborative operation of kitchen appliances. Background Technology

[0002] With the improvement of living standards and the popularization of smart homes, the intelligence level of kitchen appliances is constantly improving. Modern family kitchens are usually equipped with a variety of kitchen appliances, such as rice cookers, ovens, induction cookers, and steamers, which together form a kitchen appliance group.

[0003] During the cooking process, multiple kitchen appliances often need to work together to complete a series of complex cooking tasks. Traditional kitchen appliances mostly operate independently, requiring manual planning of the order and timing of their use to complete the cooking process. With the development of smart home technology, kitchen appliances have begun to have intelligent functions, enabling information interaction and simple collaboration between kitchen appliances through network connection. However, existing collaborative kitchen appliances usually adopt a preset fixed working mode, which still has problems such as the inability to break down the process and schedule resources, lack of ability to analyze the complex temporal dependencies between processes, inability to identify key processes and prioritize resource allocation, limited ability to execute processes in parallel, and difficulty in dynamically adjusting scheduling strategies under resource competition. Summary of the Invention

[0004] This invention provides an intelligent scheduling and control method and system for collaborative work of kitchen appliances, which can at least solve some of the problems existing in the prior art.

[0005] A first aspect of this invention provides an intelligent scheduling and control method for collaborative operation of kitchen appliance groups, comprising: The system acquires the operating status information and cooking task information of each kitchen appliance in the kitchen appliance group. It then decomposes the cooking task information into processes based on preset cooking process constraints, identifies the temporal dependencies between different processes to obtain a task execution graph, extracts the capability parameters of the kitchen appliances based on the operating status information, and matches and judges them with the resource requirement parameters of the process nodes in the task execution graph to obtain a feasibility mapping relationship. Starting from each process node, traverse backward along the temporal dependency path in the task execution graph, count the number of subsequent processes corresponding to the current process node, accumulate the longest time span from the current process node to the end of the task execution graph, and calculate the key metric based on the number of subsequent processes and the longest time span. Based on the feasibility mapping relationship and the key quantification indicators, kitchen appliances are allocated to the process nodes to obtain preliminary scheduling and allocation results. Unallocated process nodes in the preliminary scheduling and allocation results are extracted and resource release dependency information is obtained by backtracking along the temporal dependency relationship in the task execution graph. Process node combinations that do not have temporal sequence constraints with unallocated process nodes are identified to obtain parallel execution process combinations. Kitchen appliances are then allocated again to parallel execution process combinations based on the feasibility mapping relationship to generate a kitchen appliance scheduling sequence.

[0006] In one alternative implementation, The system acquires the operating status information and cooking task information of each kitchen appliance in the kitchen appliance group, and decomposes the cooking task information into processes based on preset cooking process constraints. It then identifies the temporal dependencies between different processes to obtain a task execution graph, including: The system receives a cooking task input by the user and determines the dish identifier and cooking target parameters in the cooking task. Based on the dish identifier, it queries the preset cooking process constraint rules and extracts the process sequence corresponding to the dish identifier. The process sequence is then decomposed and the operation type and resource requirement parameters corresponding to each process are identified. Based on the process flow definition in the cooking process constraint rules, the preceding and subsequent processes of each process in the process sequence are analyzed. Based on the preceding and subsequent processes, the state transmission relationship between different processes is determined and a preliminary dependency graph is constructed. Obtain the operating status information of each kitchen appliance in the kitchen appliance group, extract the equipment occupancy time and remaining processing capacity from the operating status information, traverse the process nodes in the preliminary dependency graph and extract the resource requirement parameters corresponding to each process node, compare the resource requirement parameters with the remaining processing capacity corresponding to each kitchen appliance, identify the process whose resource requirement parameters exceed the remaining processing capacity, and determine the temporal dependency relationship between different processes to obtain the task execution graph.

[0007] In one alternative implementation, Based on the operational status information, the capability parameters of the kitchen appliances are extracted and matched with the resource requirement parameters of the process nodes in the task execution diagram to obtain the feasibility mapping relationship, including: The system analyzes the operating status information to extract the equipment type identifier and historical execution record of the kitchen appliances and determines the capability parameters of each kitchen appliance. The capability parameters include a set of functional operations and performance boundary values. The system traverses the process nodes in the task execution graph and extracts the operation type and resource requirement parameters of each process node. The system matches the operation type with the set of functional operations and identifies candidate kitchen appliances that contain the operation type of the process node in the set of functional operations. Retrieve historical process nodes that the candidate kitchen appliance has performed with the same operation type as the current process node from the historical execution records, extract the historical resource requirement parameters and actual resource consumption corresponding to the historical process nodes, and perform regression fitting to obtain a regression mapping relationship. Substitute the resource requirement parameters of the current process node into the regression mapping relationship to calculate the predicted resource consumption. The feasibility of the pre-acquired preceding and subsequent processes on the candidate kitchen appliances is determined, and the total number of executable nodes is counted. A topology consistency score is calculated based on the total number of nodes. A resource satisfaction degree is calculated based on the predicted resource consumption and the performance boundary value, and a weighted sum is obtained with the topology consistency score to obtain a matching adaptation value. Candidate kitchen appliances with matching adaptation values ​​higher than a preset matching threshold are associated with process nodes to obtain a feasibility mapping relationship.

[0008] In one alternative implementation, Starting from each process node, traverse backwards along the temporal dependency path in the task execution graph, count the number of subsequent processes corresponding to the current process node, and accumulate the longest time span from the current process node to the end of the task execution graph. Based on the number of subsequent processes and the longest time span, calculate the key metrics, including: Starting from each process node, traverse backward along the time-dependent path in the task execution graph, extract the estimated execution time of each subsequent process node until the endpoint node is reached, accumulate the estimated execution time of each process node in each reachable path to obtain the path duration and determine the maximum time span, and count the number of all subsequent process nodes reachable from the current process node to obtain the number of subsequent processes. Extract a set of historical process nodes with the same operation type as the current process node from the pre-acquired historical execution records. Obtain the temporal extension depth and corresponding task completion deviation value of each historical process node and quantify the correlation to obtain the time dimension sensitive factor. Obtain the topological coverage breadth of each historical process node and corresponding task completion deviation value and quantify the correlation to obtain the breadth dimension sensitive factor. A time-series criticality metric is obtained by nonlinearly coupling the longest time span with a time-dimensional sensitivity factor, and a topological impact metric is obtained by nonlinearly coupling the number of subsequent processes with a breadth-dimensional sensitivity factor. A basic priority score is calculated based on the time-series criticality metric and the topological impact metric. The basic priority score is encoded as a particle position vector. The criticality metric is obtained by iteratively optimizing the particle position vector using a particle swarm optimization algorithm, combined with the resource requirement parameters corresponding to the current process node and the remaining capacity parameters corresponding to the kitchen appliances.

[0009] In one alternative implementation, The basic priority score is encoded as a particle position vector. Combined with the resource requirement parameters corresponding to the current process node and the remaining capacity parameters corresponding to the kitchen appliances, the key metric indicators are obtained by iteratively optimizing the particle position vector using a particle swarm optimization algorithm. These indicators include: The basic priority scores of each process node are vectorized and encoded according to the arrangement order of the process nodes in the task execution diagram to obtain the initial position vector. The constraint penalty term is determined based on the matching tension between the resource demand parameter and the remaining capacity parameter. With the goal of minimizing the task completion time, the fitness corresponding to the initial position vector is calculated as the individual optimal fitness in combination with the pre-set constraint penalty term. The particle position vector with the largest individual optimal fitness is selected as the global optimal vector. The individual cognitive component is determined based on the deviation between the particle's current position vector and the individual optimal vector, as well as a preset individual learning factor. The social cognitive component is determined based on the deviation between the particle's current position vector and the global optimal vector, as well as a preset global influence coefficient. The particle velocity is updated based on the individual cognitive component and the social cognitive component to obtain an updated velocity vector. The candidate position vector is obtained by fusing the particle's current position vector and the updated velocity vector. The candidate position vectors are tested for feasibility region. Based on the test results, new position vectors are determined. The fitness of the new position vectors is calculated and compared with the individual optimal fitness. Based on the comparison results, the individual optimal vectors and the global optimal vectors are updated. The updates are repeated until the preset maximum number of iterations is reached. The priority values ​​of each process node in the global optimal vector are extracted to obtain key metric indicators.

[0010] In one alternative implementation, Based on the feasibility mapping relationship and the key quantification indicators, kitchen appliances are allocated to the process nodes to obtain preliminary scheduling and allocation results. Unallocated process nodes in the preliminary scheduling and allocation results are extracted, and resource release dependency information is obtained by backtracking along the temporal dependency relationship in the task execution graph, including: Each process node is sorted in descending order according to key quantifiable indicators. Based on the sorting results, the set of executable kitchen appliances corresponding to each process node is extracted from the feasibility mapping relationship. Based on the matching relationship between the remaining capacity parameter and the resource demand parameter, kitchen appliances are allocated to the process nodes to obtain the preliminary scheduling and allocation results. Extract all unassigned process nodes from the preliminary scheduling and allocation results, construct a predecessor tracing path by backtracking along the temporal dependency in the task execution graph, determine the resource release propagation depth based on the allocation status of each process node, and select the allocation target device corresponding to the allocated process node with the largest resource release propagation depth as the priority candidate device for the unassigned process node. For an unassigned process node with multiple predecessor tracing paths, the priority candidate device and the corresponding resource release time point in each predecessor tracing path are extracted. The timing matching degree between the path duration and the resource release time point of each predecessor tracing path is calculated, and the path priority is determined based on the timing matching degree. The priority candidate device corresponding to the predecessor tracing path with the highest path priority is taken as the optimal candidate device for the unassigned process node. A delayed binding mapping relationship is established between the current unassigned process node and the priority candidate device and the corresponding resource release time point to obtain resource release dependency information.

[0011] In one alternative implementation, Combinations of process nodes that do not have timing constraints with unassigned process nodes are identified to obtain parallel-executable process combinations. Based on the feasibility mapping relationship, kitchen appliances are then reassigned to these parallel-executable process combinations, generating a kitchen appliance scheduling sequence including: Based on the resource release dependency information, an adjacency matrix of the task execution graph is constructed and a transitive closure matrix is ​​calculated. Row vectors and column vectors corresponding to unassigned process nodes are extracted from the transitive closure matrix. The position indices of process nodes whose element values ​​in both row and column vectors are zero are used as candidate parallel nodes. Process nodes with the same priority backup equipment and a resource release time difference less than a preset time threshold are selected from the candidate parallel nodes. The resource feasibility of the process nodes and the remaining capacity parameters of the priority backup equipment are verified by combining the feasibility mapping relationship. The process nodes that pass the verification are combined to obtain a combination of parallel executable processes. Calculate the time dispersion of resource release time points corresponding to each process node in the parallelizable process combination. Calculate resource sufficiency based on the remaining capacity parameters of priority candidate devices and the combined resource requirements corresponding to the parallelizable process combination. Calculate the temporal resource synergy based on the time dispersion and resource sufficiency. Select the priority candidate device with the highest temporal resource synergy as the target parallel device. Allocate execution time windows based on the execution duration of the process nodes and the resource release time points of the target parallel devices. Generate a kitchen appliance scheduling sequence by topologically sorting each process node, its corresponding target device, and the execution time window according to the temporal dependency relationship.

[0012] A second aspect of the present invention provides an intelligent scheduling and control system for collaborative operation of kitchen appliances, comprising: The first unit is used to acquire the operating status information and cooking task information of each kitchen appliance in the kitchen appliance group, and to decompose the cooking task information to be executed into processes in combination with preset cooking process constraint rules, identify the temporal dependency relationship between different processes to obtain a task execution diagram, extract the capability parameters of the kitchen appliances based on the operating status information and match them with the resource requirement parameters of the process nodes in the task execution diagram to obtain a feasibility mapping relationship. The second unit is used to traverse backward along the temporal dependency path in the task execution graph starting from each process node, count the number of subsequent processes corresponding to the current process node, accumulate the longest time span from the current process node to the end of the task execution graph, and calculate key metric indicators based on the number of subsequent processes and the longest time span. The third unit is used to allocate kitchen appliances to the process nodes according to the feasibility mapping relationship and the key quantification index to obtain a preliminary scheduling allocation result, extract the unallocated process nodes in the preliminary scheduling allocation result and backtrack along the temporal dependency relationship in the task execution graph to obtain resource release dependency information, identify process node combinations that do not have temporal order constraints with the unallocated process nodes to obtain parallel executable process combinations, and re-allocate kitchen appliances to the parallel executable process combinations in combination with the feasibility mapping relationship to generate a kitchen appliance scheduling sequence.

[0013] A third aspect of the present invention provides an electronic device, comprising: A processor and a memory for storing processor-executable instructions, wherein the processor is configured to invoke instructions stored in the memory to perform the aforementioned method.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0015] In this invention, by acquiring the operating status and cooking task information of each device in the kitchen appliance group, and combining the cooking process constraint rules to decompose the process, the temporal dependencies between different processes can be accurately identified, enabling refined management of complex cooking tasks. By analyzing the number of subsequent processes and the longest time span of each process node, the importance of the process can be quantitatively evaluated, making scheduling decisions more scientific and reasonable, avoiding potential uneven resource allocation and task delays, and effectively improving overall cooking efficiency. By identifying parallel process combinations and performing secondary resource allocation, efficient allocation of kitchen appliance resources is achieved. Attached Figure Description

[0016] Figure 1This is a flowchart illustrating the intelligent scheduling and control method for collaborative operation of kitchen appliances in an embodiment of the present invention. Figure 2 This is a flowchart illustrating the time-series resource collaborative scheduling process of the intelligent scheduling and control method for collaborative work of kitchen appliance groups according to an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0019] Figure 1 This is a flowchart illustrating the intelligent scheduling and control method for collaborative operation of kitchen appliances in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes: The system acquires the operating status information and cooking task information of each kitchen appliance in the kitchen appliance group. It then decomposes the cooking task information into processes based on preset cooking process constraints, identifies the temporal dependencies between different processes to obtain a task execution graph, extracts the capability parameters of the kitchen appliances based on the operating status information, and matches and judges them with the resource requirement parameters of the process nodes in the task execution graph to obtain a feasibility mapping relationship. Starting from each process node, traverse backward along the temporal dependency path in the task execution graph, count the number of subsequent processes corresponding to the current process node, accumulate the longest time span from the current process node to the end of the task execution graph, and calculate the key metric based on the number of subsequent processes and the longest time span. Based on the feasibility mapping relationship and the key quantification indicators, kitchen appliances are allocated to the process nodes to obtain preliminary scheduling and allocation results. Unallocated process nodes in the preliminary scheduling and allocation results are extracted and resource release dependency information is obtained by backtracking along the temporal dependency relationship in the task execution graph. Process node combinations that do not have temporal sequence constraints with unallocated process nodes are identified to obtain parallel execution process combinations. Kitchen appliances are then allocated again to parallel execution process combinations based on the feasibility mapping relationship to generate a kitchen appliance scheduling sequence.

[0020] In one alternative implementation, The system acquires the operating status information and cooking task information of each kitchen appliance in the kitchen appliance group, and decomposes the cooking task information into processes based on preset cooking process constraints. It then identifies the temporal dependencies between different processes to obtain a task execution graph, including: The system receives a cooking task input by the user and determines the dish identifier and cooking target parameters in the cooking task. Based on the dish identifier, it queries the preset cooking process constraint rules and extracts the process sequence corresponding to the dish identifier. The process sequence is then decomposed and the operation type and resource requirement parameters corresponding to each process are identified. Based on the process flow definition in the cooking process constraint rules, the preceding and subsequent processes of each process in the process sequence are analyzed. Based on the preceding and subsequent processes, the state transmission relationship between different processes is determined and a preliminary dependency graph is constructed. Obtain the operating status information of each kitchen appliance in the kitchen appliance group, extract the equipment occupancy time and remaining processing capacity from the operating status information, traverse the process nodes in the preliminary dependency graph and extract the resource requirement parameters corresponding to each process node, compare the resource requirement parameters with the remaining processing capacity corresponding to each kitchen appliance, identify the process whose resource requirement parameters exceed the remaining processing capacity, and determine the temporal dependency relationship between different processes to obtain the task execution graph.

[0021] The system receives user input for a cooking task and uses natural language processing (NLP) to identify dish identifiers and cooking target parameters. For example, when a user inputs "make a braised pork rib with a rich caramel color," it identifies "braised pork rib" as the dish identifier and "rich caramel color" as the cooking target parameter. A matching algorithm based on a dish lexicon is used to segment the user input and match it with dish names in a pre-defined lexicon. Matches with a similarity score exceeding 0.85 are identified as dish identifiers. Simultaneously, cooking-related adjectives and adverbs are extracted as cooking target parameters.

[0022] Based on the identified dish identifier, the corresponding process sequence is retrieved from a pre-defined cooking process constraint rule library. This library stores rules in key-value pairs, where the key is the dish identifier and the value is the process sequence containing the order of operations. For "Braised Pork Ribs," the retrieved process sequence includes steps such as "ingredient preparation → blanching meat → pan-frying and browning → stewing → seasoning and reducing sauce." Each step is further broken down to identify its operation type and resource requirements. Operation types include cutting, frying, steaming, and stewing; resource requirements include equipment type, power requirements, container size, temperature range, and duration. For example, the operation type for "pan-frying and browning" is frying, with resource requirements of a medium heat temperature range of 165-180℃, a duration of 3-5 minutes, an induction cooker or gas stove, and a flat-bottomed frying pan.

[0023] Based on the process flow definition in the cooking process constraint rules, the predecessor and successor processes of each process in the process sequence are analyzed. By traversing the process sequence, a list of predecessor and successor processes is established for each process. For example, the predecessor process of "frying and browning" is "blanching meat," and the successor process is "stewing." Based on the aforementioned relationships, a preliminary dependency graph is constructed. The preliminary dependency graph adopts a directed acyclic graph structure, where nodes represent processes and edges represent dependencies between processes. For each edge in the dependency graph, it indicates that the starting process must be completed before the terminating process. This graph structure is stored using an adjacency list or adjacency matrix to ensure that the process execution order conforms to the cooking process requirements.

[0024] The system acquires the operating status information of each kitchen appliance in the kitchen appliance group, including appliance type, current operating mode, temperature, remaining time, occupied time period, and remaining processing capacity. This data is obtained in real-time via the device's network interface and stored in a status cache indexed by the device number. For example, a rice cooker currently in keep-warm mode at 65℃ can be immediately switched to another operating mode; an induction cooker is currently operating at 180℃ and is expected to complete its current task in 15 minutes. The system extracts the occupied time period and remaining processing capacity, which is represented by the difference between the device's maximum power and the currently used power.

[0025] Traverse the process nodes in the initial dependency graph and extract the resource requirement parameters corresponding to each process node. For each process, select a subset of kitchen appliances that meet its resource requirements and compare the resource requirement parameters of the process with the remaining processing capacity of each candidate appliance. When the resource requirement parameters of a process exceed the remaining processing capacity of all available appliances, the process is marked as a resource-constrained process. For example, the "stewing" process requires continuous heating for 1.5 hours, maintaining a temperature of 95-98℃, but all electric slow cookers are currently occupied, and it is estimated that only one appliance will be available after 30 minutes. In this case, the "stewing" process is marked as a resource-constrained process, and a resource waiting time of 30 minutes is set.

[0026] Based on the resource-constrained analysis, the temporal dependencies between different processes are further determined. In addition to the predecessor-successor relationships defined in the process flow, temporal dependencies caused by resource competition are also added. When two processes that could originally run in parallel need to use the same equipment, and that equipment can only execute one process at a time, a temporal dependency edge is added between these two processes, with the higher-priority process executing first. Priority is determined by a combination of factors, including the urgency of the process and its impact on the final cooking result. A task execution graph is constructed, containing the complete execution sequence and time arrangement of the processes.

[0027] In this embodiment, by structurally breaking down the processes in the cooking task and identifying the operation types and resource requirements, the cooking process can be understood from a fine-grained perspective, avoiding the problems of rough scheduling and resource waste caused by executing the entire dish. The preliminary dependency graph constructed based on the process flow definition can accurately depict the logical relationship between processes, improve the rationality of task flow, and avoid execution conflicts or delays caused by unclear dependencies. By combining the real-time operating status of kitchen appliances with the remaining processing capacity to match process resources, steps with insufficient resources can be identified in advance and the temporal dependencies can be clarified, effectively reducing the probability of equipment contention, queuing blockage, and execution conflicts.

[0028] In one alternative implementation, Based on the operational status information, the capability parameters of the kitchen appliances are extracted and matched with the resource requirement parameters of the process nodes in the task execution diagram to obtain the feasibility mapping relationship, including: The system analyzes the operating status information to extract the equipment type identifier and historical execution record of the kitchen appliances and determines the capability parameters of each kitchen appliance. The capability parameters include a set of functional operations and performance boundary values. The system traverses the process nodes in the task execution graph and extracts the operation type and resource requirement parameters of each process node. The system matches the operation type with the set of functional operations and identifies candidate kitchen appliances that contain the operation type of the process node in the set of functional operations. Retrieve historical process nodes that the candidate kitchen appliance has performed with the same operation type as the current process node from the historical execution records, extract the historical resource requirement parameters and actual resource consumption corresponding to the historical process nodes, and perform regression fitting to obtain a regression mapping relationship. Substitute the resource requirement parameters of the current process node into the regression mapping relationship to calculate the predicted resource consumption. The feasibility of the pre-acquired preceding and subsequent processes on the candidate kitchen appliances is determined, and the total number of executable nodes is counted. A topology consistency score is calculated based on the total number of nodes. A resource satisfaction degree is calculated based on the predicted resource consumption and the performance boundary value, and a weighted sum is obtained with the topology consistency score to obtain a matching adaptation value. Candidate kitchen appliances with matching adaptation values ​​higher than a preset matching threshold are associated with process nodes to obtain a feasibility mapping relationship.

[0029] The acquired operational status information is analyzed to extract the equipment type identifiers and historical execution records of the kitchen appliances, and the capability parameters of each appliance are determined. Equipment type identifiers include induction cookers, electric steamers, electric slow cookers, ovens, etc.; historical execution records include information on the processes performed by the equipment, execution time, and resource consumption. Capability parameters mainly include two parts: the set of functional operations and performance boundary values. The set of functional operations describes the types of operations the equipment can perform, such as frying, boiling, and blanching for an induction cooker; performance boundary values ​​describe the upper and lower limits of the equipment's performance, such as a temperature range of 60-280℃, a power range of 200-2000 watts, and a size limit of 30 cm in diameter for an induction cooker. For a multi-functional electric pressure cooker, its set of functional operations may include cooking rice, braising, stewing, and steaming, and its performance boundary values ​​include a maximum pressure of 70 kPa, a temperature range of 40-110℃, and a capacity of 5 liters.

[0030] The process iterates through the process nodes in the task execution graph, extracting the operation type and resource requirement parameters for each node. The operation type is then matched against the functional operation sets of each kitchen appliance, identifying candidate appliances whose functional operation sets contain the operation type of the process node. For example, for the process node with the operation type "frying and browning," the resource requirement parameters are a temperature of 170℃ and a duration of 4 minutes. Matching reveals that the functional operation sets of induction cookers and electric griddles contain this operation type, thus marking them as candidate appliances. The matching process uses a semantic similarity calculation method, converting the operation type into a feature vector and calculating the similarity between the feature vector and each operation type in the functional operation set. A similarity score greater than 0.85 is considered a successful match.

[0031] The system retrieves historical process nodes from the execution records that have been performed by candidate kitchen appliances with the same operation type as the current process node. It extracts the historical resource requirement parameters and actual resource consumption corresponding to these historical process nodes, and performs regression fitting to obtain a regression mapping relationship. This regression mapping relationship reflects the correspondence between resource requirement parameters and actual resource consumption, and is established using methods such as multiple linear regression or decision tree regression. Taking the "frying and browning" operation performed by an induction cooker as an example, 20 historical records are extracted. Historical resource requirement parameters include set temperature, duration, and ingredient weight, while actual resource consumption includes actual power consumption and actual execution time. Through multiple linear regression analysis, the regression mapping relationship is obtained: actual power consumption is positively correlated with set temperature, duration, and ingredient weight, with correlation coefficients of 0.72, 0.85, and 0.63, respectively. Substituting the resource requirement parameters of the current process node into the regression mapping relationship, the predicted resource consumption is calculated. If the current "frying and coloring" process is set to a temperature of 170℃ and a duration of 4 minutes, and the food weight is 500 grams, then by substituting the regression mapping relationship, the predicted power consumption is 0.15 kWh, and the predicted actual execution time is 4.5 minutes.

[0032] The feasibility of pre-acquired predecessor and successor processes on candidate kitchen appliances is determined, and the total number of executable nodes is counted. For each candidate kitchen appliance, it is checked whether the predecessor and successor processes of the current process node can also be executed on that appliance. The feasibility check is also achieved through functional operation set matching. The total number of predecessor and successor process nodes that can be executed by the candidate appliance is counted, and the topology consistency score is calculated. The topology consistency score reflects the frequency of process transfer between appliances; the higher the score, the less process transfer and the higher the collaboration efficiency. The topology consistency score is calculated as: the number of executable nodes divided by the total number of nodes (the sum of the current process and its predecessor and successor processes). If the induction cooker can execute the "frying and coloring" process and two of its predecessor processes "blanching" and successor processes "stewing", with a total of 3 nodes, then the topology consistency score is 2 / 3, approximately 0.67.

[0033] Resource satisfaction is calculated based on predicted resource consumption and performance boundary values. Resource satisfaction reflects the degree to which equipment capacity meets the resource requirements of a process, typically considering multiple dimensions such as temperature, power, and capacity. For each dimension, the predicted resource consumption is compared with the performance boundary value. If it is within the performance boundary, the satisfaction level for that dimension is 1; otherwise, it is scored between 0 and 1 according to the degree to which it exceeds the boundary. The weighted average of the satisfaction levels for each dimension yields the overall resource satisfaction. If the induction cooker has a temperature satisfaction level of 1 for the "frying and coloring" process (i.e., a temperature of 170℃ within the 60-280℃ range), a power satisfaction level of 1 (i.e., a power of 1200 watts within the 200-2000 watt range), and a size satisfaction level of 0.9 (slightly exceeding the limit), then the overall resource satisfaction level is 0.97.

[0034] The matching fit value is obtained by weighted summation of resource satisfaction and topology consistency score. The matching fit value comprehensively reflects the suitability of kitchen appliances for performing a specific process. In the weighted summation formula, resource satisfaction typically has a weight of 0.7, and topology consistency score has a weight of 0.3. For an induction cooker performing the "frying and browning" process, the matching fit value is 0.97 × 0.7 + 0.67 × 0.3 = 0.88. Candidate kitchen appliances with matching fit values ​​higher than a preset matching threshold are associated with the process node to obtain a feasibility mapping relationship. The preset matching threshold is typically set to 0.75. The feasibility mapping relationship is stored in key-value pairs, where the key is the process node identifier, and the value is a list of kitchen appliances capable of performing the process and their corresponding matching fit values.

[0035] In this embodiment, by parsing the operating status information and extracting the equipment type identifier, functional operation set, and performance boundary value, resources can be meticulously characterized from both the equipment function and performance levels. This overcomes the shortcomings of coarse-grained matching based on static equipment categories or simple capacity indicators, significantly improving the accuracy of equipment selection. By constructing a regression mapping relationship based on historical execution records and using the mapping relationship to predict the resource consumption of the current process, resource estimation bias is reduced and the reliability of matching judgment is improved. By introducing the executability analysis of the predecessor and successor processes on the equipment, a topology consistency score is formed, which can ensure that the process scheduling maintains consistency in logical order, avoid cross-equipment scheduling interruptions, frequent switching, or scheduling deadlocks caused by ignoring process continuity, and improve the overall process coherence.

[0036] In one alternative implementation, Starting from each process node, traverse backwards along the temporal dependency path in the task execution graph, count the number of subsequent processes corresponding to the current process node, and accumulate the longest time span from the current process node to the end of the task execution graph. Based on the number of subsequent processes and the longest time span, calculate the key metrics, including: Starting from each process node, traverse backward along the time-dependent path in the task execution graph, extract the estimated execution time of each subsequent process node until the endpoint node is reached, accumulate the estimated execution time of each process node in each reachable path to obtain the path duration and determine the maximum time span, and count the number of all subsequent process nodes reachable from the current process node to obtain the number of subsequent processes. Extract a set of historical process nodes with the same operation type as the current process node from the pre-acquired historical execution records. Obtain the temporal extension depth and corresponding task completion deviation value of each historical process node and quantify the correlation to obtain the time dimension sensitive factor. Obtain the topological coverage breadth of each historical process node and corresponding task completion deviation value and quantify the correlation to obtain the breadth dimension sensitive factor. A time-series criticality metric is obtained by nonlinearly coupling the longest time span with a time-dimensional sensitivity factor, and a topological impact metric is obtained by nonlinearly coupling the number of subsequent processes with a breadth-dimensional sensitivity factor. A basic priority score is calculated based on the time-series criticality metric and the topological impact metric. The basic priority score is encoded as a particle position vector. The criticality metric is obtained by iteratively optimizing the particle position vector using a particle swarm optimization algorithm, combined with the resource requirement parameters corresponding to the current process node and the remaining capacity parameters corresponding to the kitchen appliances.

[0037] Starting from each process node, the process traverses backward along the temporal dependency path in the task execution graph, extracting the estimated execution time of each subsequent process node until the endpoint is reached. The estimated execution time is calculated based on historical data and resource requirement parameters. The estimated execution time of each process node in each reachable path is summed to obtain the path duration, and the longest time span is determined. The longest time span represents the longest path time from the current process node to the endpoint, reflecting the degree of influence of this process on the overall completion time of the cooking task. Taking the braised pork ribs cooking task as an example, starting from the "pan-frying and browning" process node, the reachable paths include "pan-frying and browning → stewing → seasoning and reducing sauce" and "pan-frying and browning → stewing → cooling". Assuming that the estimated execution time of "pan-frying and browning" is 5 minutes, "stewing" is 60 minutes, "seasoning and reducing sauce" is 10 minutes, and "cooling" is 15 minutes, then the first path duration is 75 minutes, the second is 80 minutes, and the longest time span is 80 minutes. Simultaneously, the number of all subsequent process nodes reachable from the current process node is counted to obtain the number of subsequent processes. In the example above, the number of subsequent processes for the "frying and coloring" process is 3, including "stewing", "seasoning and reducing sauce" and "cooling".

[0038] The system extracts a set of historical process nodes with the same operation type as the current process node from pre-acquired historical execution records. These historical execution records store detailed execution information for each process in past completed cooking tasks. For each historical process node, the corresponding temporal extension depth and task completion deviation are obtained. Temporal extension depth refers to the longest path length from the current process to the final process, expressed as the number of processes; task completion deviation refers to the ratio of the actual completion time to the expected completion time. By quantitatively analyzing the correlation between these two indicators, a time-dimensional sensitivity factor is obtained. This time-dimensional sensitivity factor reflects the degree of influence of the process's temporal position on the overall task completion time. The quantitative analysis uses the Pearson correlation coefficient. Assuming an analysis of 20 historical records, the time-dimensional sensitivity factor for the "frying and browning" process is 0.78, indicating a strong positive correlation between the temporal extension depth and the task completion deviation value, meaning that the delay of this process has a significant impact on the overall task completion time.

[0039] The topology coverage breadth and corresponding task completion deviation value of each historical process node are obtained. Topology coverage breadth refers to the proportion of subsequent processes directly or indirectly affected by this process out of the total number of processes. By quantitatively analyzing the correlation between these two indicators, a breadth dimension sensitivity factor is obtained. The breadth dimension sensitivity factor reflects the degree of influence of the process topology location on the overall task completion. Assuming the same 20 historical records are analyzed, the breadth dimension sensitivity factor of the "frying and coloring" process is 0.65, indicating that the topology coverage breadth of this process is moderately positively correlated with the task completion deviation value, that is, the quality fluctuation of this process will have a chain reaction effect on many subsequent processes.

[0040] A nonlinear coupling was performed between the longest time span and the time-dimensional sensitivity factor to obtain a time-series criticality metric. The nonlinear coupling used a logarithmically weighted product form, where the time-series criticality metric equals the product of the natural logarithm of the longest time span and the time-dimensional sensitivity factor, plus a base weight (usually 1). For the "frying and coloring" process, the longest time span is 80 minutes, the time-dimensional sensitivity factor is 0.78, and the calculated time-series criticality metric is 4.38. A similar nonlinear coupling was performed between the number of subsequent processes and the breadth-dimensional sensitivity factor to obtain a topological influence metric. For the "frying and coloring" process, the number of subsequent processes is 3, the breadth-dimensional sensitivity factor is 0.65, and the calculated topological influence metric is 2.14.

[0041] The basic priority score is calculated based on temporal criticality and topological influence metrics. A weighted average method is used for the basic priority score calculation, with a weight of 0.6 for temporal criticality and 0.4 for topological influence. For the "frying and browning" process, the basic priority score is 4.38 × 0.6 + 2.14 × 0.4 = 3.48. The basic priority score is encoded as a particle position vector, with each particle representing a priority allocation scheme for a process. The encoding method involves normalizing the basic priority score to the 0-1 interval, which serves as the particle's position value in the first dimension; simultaneously, the resource requirement parameters corresponding to the current process node and the remaining capacity parameters corresponding to the kitchen appliances are considered and encoded as the particle's position values ​​in other dimensions.

[0042] The particle swarm optimization (PSO) algorithm iteratively optimizes the particle position vectors to obtain key metrics. The PSO algorithm finds the optimal solution by continuously updating the particle positions and velocities. The process involves initializing the particle swarm, with each particle randomly distributed in the search space; defining a fitness function that considers both the rationality of process priorities and the efficiency of resource allocation, with higher fitness values ​​indicating better allocation schemes; iteratively updating particle positions and velocities, with each iteration updating the movement direction and distance based on its historical best position and the global best position; and stopping the algorithm when the maximum number of iterations is reached or the convergence condition is met, outputting the globally optimal particle position, i.e., the key metric.

[0043] In this embodiment, by traversing backward along the temporal dependency path of the task execution graph and calculating the longest time span and the number of subsequent processes, the importance of the process in the overall task chain can be identified from a global process perspective. This avoids scheduling based on single node attributes or local features and significantly improves the accuracy of key process identification. By quantitatively correlating the execution deviation values ​​of historical process nodes with temporal extension depth and topological coverage breadth, and establishing time-dimensional and breadth-dimensional sensitive factors, the impact of processes on the overall task achievement quality can be evaluated based on actual execution performance. This overcomes the limitations of insufficient utilization of historical data and inability to quantify process influence, making key metrics more consistent with actual operating rules. By introducing a particle swarm optimization algorithm to optimize the basic priority score, the stability of scheduling results and resource utilization efficiency can be significantly improved.

[0044] In one alternative implementation, The basic priority score is encoded as a particle position vector. Combined with the resource requirement parameters corresponding to the current process node and the remaining capacity parameters corresponding to the kitchen appliances, the key metric indicators are obtained by iteratively optimizing the particle position vector using a particle swarm optimization algorithm. These indicators include: The basic priority scores of each process node are vectorized and encoded according to the arrangement order of the process nodes in the task execution diagram to obtain the initial position vector. The constraint penalty term is determined based on the matching tension between the resource demand parameter and the remaining capacity parameter. With the goal of minimizing the task completion time, the fitness corresponding to the initial position vector is calculated as the individual optimal fitness in combination with the pre-set constraint penalty term. The particle position vector with the largest individual optimal fitness is selected as the global optimal vector. The individual cognitive component is determined based on the deviation between the particle's current position vector and the individual optimal vector, as well as a preset individual learning factor. The social cognitive component is determined based on the deviation between the particle's current position vector and the global optimal vector, as well as a preset global influence coefficient. The particle velocity is updated based on the individual cognitive component and the social cognitive component to obtain an updated velocity vector. The candidate position vector is obtained by fusing the particle's current position vector and the updated velocity vector. The candidate position vectors are tested for feasibility region. Based on the test results, new position vectors are determined. The fitness of the new position vectors is calculated and compared with the individual optimal fitness. Based on the comparison results, the individual optimal vectors and the global optimal vectors are updated. The updates are repeated until the preset maximum number of iterations is reached. The priority values ​​of each process node in the global optimal vector are extracted to obtain key metric indicators.

[0045] The basic priority scores of each process node are vectorized according to their arrangement in the task execution graph to obtain the initial position vector. Taking a braised pork ribs cooking task with 5 processes as an example, the process nodes and their corresponding basic priority scores are as follows: ingredient preparation 2.45, meat blanching 3.12, frying and browning 3.48, stewing 3.95, and seasoning and reducing sauce 3.08. The basic priority scores are normalized to obtain the initial position vector [0.52, 0.66, 0.74, 0.84, 0.65]. Each component of the vector ranges from 0 to 1, with a larger value indicating a higher priority. When initializing the particle swarm, to reduce the impact of randomness on the optimization effect, the basic priority of each process is used as a reference value for the initial position of the particles, while a small random perturbation is introduced. For example, for the initialization of 10 particles, the initial position vector of the first particle may be [0.53, 0.67, 0.73, 0.85, 0.64], and other particles are similarly initialized with random perturbation.

[0046] Constraint penalties are determined based on the matching tension between resource demand parameters and remaining capacity parameters. Matching tension reflects the degree of competition for resources by a process, calculated as the ratio of resource demand to remaining equipment capacity. When demand exceeds remaining capacity, the matching tension is greater than 1, indicating insufficient resources; when demand is less than remaining capacity, the matching tension is less than 1, indicating sufficient resources. The principle for setting constraint penalties is: the higher the matching tension, the heavier the penalty. Specifically, when the matching tension is less than 1, the constraint penalty is 0; when the matching tension is greater than or equal to 1, the constraint penalty is the matching tension minus 1 multiplied by a penalty coefficient, typically set to 100. For example, the frying and coloring process requires an induction cooker with a temperature of 170℃ and a power of 1200 watts, while the current remaining capacity of the induction cooker is 250℃ and a power of 1000 watts. Therefore, the matching tension for the temperature dimension is 170 / 250 = 0.68, and the matching tension for the power dimension is 1200 / 1000 = 1.2. The temperature dimension constraint penalty is 0, and the power dimension constraint penalty is 20.

[0047] The goal is to minimize task completion time. The fitness of the initial position vector is calculated using pre-set constraint penalties, and this fitness is used as the individual's optimal fitness. The fitness function is designed as the reciprocal of the total task completion time minus the constraint penalties. The total task completion time is calculated using the critical path, i.e., finding the longest path from the starting process to the ending process under the current priority allocation scheme. The constraint penalties are the sum of the constraints for each process. A higher fitness value indicates a better allocation scheme. For the initial position vector [0.53, 0.67, 0.73, 0.85, 0.64], assuming the calculated total task completion time is 95 minutes and the total constraint penalties are 30, the fitness value is 1 / 95 - 30 / 10000 = -0.0097. The particle position vector with the highest individual optimal fitness is selected as the global optimal vector. Assuming that among the initial 10 particles, the third particle has the highest fitness value of 0.0086, the position vector of the third particle is set as the global optimal vector.

[0048] The individual cognitive component is determined based on the deviation between the particle's current position vector and its optimal vector, along with a preset individual learning factor. The individual cognitive component represents the particle's tendency to move towards its historical best position. It is calculated by subtracting the particle's current position vector from its optimal vector, multiplying by the individual learning factor (usually set to 2.0), and then adding a random number between 0 and 1. For a particle with current position vectors of [0.53, 0.67, 0.73, 0.85, 0.64], optimal vectors of [0.55, 0.68, 0.72, 0.86, 0.63], an individual learning factor of 2.0, and a randomly generated number of 0.6, the individual cognitive component would be [0.024, 0.012, -0.012, 0.012, -0.012].

[0049] The social cognitive component is determined based on the deviation between the particle's current position vector and the global optimal vector, as well as a preset global influence coefficient. The social cognitive component represents the particle's tendency to move towards the optimal position within the group. It is calculated by subtracting the particle's current position vector from the global optimal vector, multiplying by the global influence coefficient (usually set to 2.0), and then adding a random number between 0 and 1. For example, assuming the global optimal vector is [0.54, 0.69, 0.75, 0.83, 0.66], the global influence coefficient is 2.0, and the randomly generated number is 0.7, then the social cognitive component is [0.014, 0.028, 0.028, -0.028, 0.028].

[0050] The updated velocity vector is obtained by updating the particle velocity based on the individual cognitive component and the social cognitive component. The update method is to multiply the current velocity vector by the inertia weight, and then add the individual cognitive component and the social cognitive component. The inertia weight is usually set to 0.8, which is used to control the degree to which the particle maintains its original motion trend. For the initial velocity vector [0.01, -0.02, 0.01, -0.01, 0.02], the updated velocity vector is [0.01×0.8+0.024+0.014, -0.02×0.8+0.012+0.028, ...], that is, [0.046, 0.024, 0.022, -0.022, 0.032].

[0051] The candidate position vector is obtained by fusing the particle's current position vector and update velocity vector. The fusion method is to add the update velocity vector to the current position vector. For the current position vector [0.53, 0.67, 0.73, 0.85, 0.64] and the update velocity vector [0.046, 0.024, 0.022, -0.022, 0.032], the fused candidate position vector is [0.576, 0.694, 0.752, 0.828, 0.672].

[0052] The candidate position vectors are subjected to a feasible region check, and the new position vectors are determined based on the check results. The feasible region check mainly includes two parts: boundary check and conflict check. The boundary check ensures that the values ​​of each component are within the range of 0 to 1. If they exceed the range, they are truncated to the boundary value. The conflict check ensures that the priorities of different processes are not too close. It is usually required that the priority difference between any two processes is not less than 0.05. If the condition is not met, the values ​​of each component are adjusted appropriately. For example, the boundary check of the candidate position vector [0.576, 0.694, 0.752, 0.828, 0.672] shows that all components are within the range of 0 to 1 and no adjustment is needed. The conflict check shows that the difference between the second component 0.694 and the fifth component 0.672 is 0.022, which is less than 0.05, and it needs to be adjusted to [0.576, 0.70, 0.752, 0.828, 0.65].

[0053] Calculate the fitness corresponding to the new position vector and compare it with the individual's optimal fitness. Update the individual's optimal vector and the global optimal vector based on the comparison result. If the fitness of the new position vector is higher than the individual's optimal fitness, set the new position vector as the individual's optimal vector and update the individual's optimal fitness; otherwise, leave it unchanged. For all particles, if any particle's individual optimal fitness is higher than the current global optimal fitness, update the global optimal vector and the global optimal fitness. Assuming the fitness of the new position vector [0.576, 0.70, 0.752, 0.828, 0.65] is 0.0102, which is higher than the individual's optimal fitness of 0.0086, update the individual's optimal vector to the new position vector and update the individual's optimal fitness to 0.0102; since 0.0102 is higher than the global optimal fitness of 0.0086, the global optimal vector and the global optimal fitness are also updated.

[0054] The update process is repeated until the preset maximum number of iterations is reached, typically set to 100. After each iteration, the inertia weight can be linearly decreased to enhance the convergence ability in later stages. For example, the initial inertia weight is 0.8, which eventually decreases to 0.4. After 100 iterations, assuming the final global optimal vector is [0.58, 0.71, 0.79, 0.87, 0.65], the corresponding global optimal fitness is 0.0128, indicating that under this priority allocation scheme, the task completion time is minimized and the resource constraint satisfaction is maximized.

[0055] The priority values ​​corresponding to each process node in the global optimal vector are extracted to obtain key metrics. Key metrics directly reflect the importance of each process in the scheduling process; the larger the value, the higher the priority. For example, for the aforementioned braised pork ribs cooking task, the final key metrics are: ingredient preparation 0.58, meat blanching 0.71, pan-frying and browning 0.79, stewing 0.87, seasoning and reducing sauce 0.65.

[0056] In this embodiment, the basic priority score is vectorized and encoded according to the process sequence of the task execution graph. A matching tension penalty term is constructed by combining resource requirement parameters and equipment remaining capacity parameters. This comprehensively considers process logic and resource feasibility, overcoming the limitations of sorting only by process factors or ignoring resource bottlenecks. This makes the starting point of the optimization process more reasonable. By defining individual cognitive components and social cognitive components, the deviation between the particle's current position and its historical best and global best is incorporated into the speed update mechanism. This prevents priority iteration from being stuck in fixed rules or local greedy strategies, which can significantly improve the convergence speed and globality of priority optimization and reduce the problem of low overall task execution efficiency caused by local optima. By performing feasible region verification on candidate position vectors, it is ensured that the new priority arrangement obtained in each iteration satisfies resource constraints and process restrictions, avoiding invalid solutions that violate dependencies or resource conflicts during the optimization process. This can improve the stability and executability of scheduling results.

[0057] In one alternative implementation, Based on the feasibility mapping relationship and the key quantification indicators, kitchen appliances are allocated to the process nodes to obtain preliminary scheduling and allocation results. Unallocated process nodes in the preliminary scheduling and allocation results are extracted, and resource release dependency information is obtained by backtracking along the temporal dependency relationship in the task execution graph, including: Each process node is sorted in descending order according to key quantifiable indicators. Based on the sorting results, the set of executable kitchen appliances corresponding to each process node is extracted from the feasibility mapping relationship. Based on the matching relationship between the remaining capacity parameter and the resource demand parameter, kitchen appliances are allocated to the process nodes to obtain the preliminary scheduling and allocation results. Extract all unassigned process nodes from the preliminary scheduling and allocation results, construct a predecessor tracing path by backtracking along the temporal dependency in the task execution graph, determine the resource release propagation depth based on the allocation status of each process node, and select the allocation target device corresponding to the allocated process node with the largest resource release propagation depth as the priority candidate device for the unassigned process node. For an unassigned process node with multiple predecessor tracing paths, the priority candidate device and the corresponding resource release time point in each predecessor tracing path are extracted. The timing matching degree between the path duration and the resource release time point of each predecessor tracing path is calculated, and the path priority is determined based on the timing matching degree. The priority candidate device corresponding to the predecessor tracing path with the highest path priority is taken as the optimal candidate device for the unassigned process node. A delayed binding mapping relationship is established between the current unassigned process node and the priority candidate device and the corresponding resource release time point to obtain resource release dependency information.

[0058] The process nodes are arranged in descending order according to their key metric values ​​to form a process priority sequence. Taking the braised pork ribs cooking task as an example, the key metric values ​​for the five process nodes are: stewing 0.87, pan-frying and browning 0.79, blanching meat 0.71, seasoning and reducing sauce 0.65, and ingredient preparation 0.58. Arranging them in descending order of metric values, the ranking result is: stewing, pan-frying and browning, blanching meat, seasoning and reducing sauce, and ingredient preparation. The ranking results show that the stewing process has the highest priority and should be allocated resources first, while the ingredient preparation process has the lowest priority.

[0059] Based on the sorting results, the set of executable kitchen appliances corresponding to each process node is extracted sequentially from the feasibility mapping relationship. The feasibility mapping relationship is a pre-established correspondence table between processes and applicable kitchen appliances, recording which devices can be used to perform each process. For example, the set of executable devices for the stewing process is {electric pressure cooker, induction cooker}, the set of executable devices for the frying and browning process is {induction cooker, electric oven}, the set of executable devices for the meat blanching process is {induction cooker}, the set of executable devices for the seasoning and sauce reduction process is {induction cooker}, and the set of executable devices for the ingredient preparation process is {food processor, vegetable cutter}.

[0060] Based on the matching relationship between remaining capacity parameters and resource demand parameters, kitchen appliances are allocated to process nodes to obtain preliminary scheduling and allocation results. Remaining capacity parameters refer to the amount of various resources currently available from the equipment, including temperature range, power, and capacity; resource demand parameters refer to the amount of various resources required for process execution. The matching relationship is evaluated using resource matching degree calculation, which is the ratio of demand parameters to remaining capacity parameters. The closer the ratio is to 1, the higher the matching degree. For the stewing process, a temperature of 100℃, a power of 800 watts, and a time of 45 minutes are required. The remaining capacity of the electric pressure cooker is 120℃ and 900 watts; the remaining capacity of the induction cooker is 250℃ and 1000 watts. The resource matching degree is calculated as follows: the temperature matching degree of the electric pressure cooker is 100 / 120 = 0.83, and the power matching degree is 800 / 900 = 0.89; the temperature matching degree of the induction cooker is 100 / 250 = 0.4, and the power matching degree is 800 / 1000 = 0.8. Taking all factors into consideration, the electric pressure cooker has a higher overall compatibility, therefore the stewing process is assigned to it. Similarly, the pan-frying and browning process is assigned to the induction cooker. The meat blanching process cannot be assigned at this time because the induction cooker is already in use. The seasoning and sauce reduction process also cannot be assigned. The ingredient preparation process is assigned to the food processor. The preliminary scheduling result is: Stewing - Electric Pressure Cooker, Pan-frying and Browning - Induction Cooker, Ingredient Preparation - Food Processor, Meat Blanching and Seasoning / Sauce Reduction are not assigned at this time.

[0061] Extract all unassigned process nodes from the initial scheduling and allocation results; in this example, these are the meat blanching and seasoning / reducing processes. Construct a predecessor tracing path by backtracking along the temporal dependencies in the task execution graph. The task execution graph is a directed graph describing the dependencies between processes; nodes represent processes, and edges represent dependencies. The predecessor tracing path is the path that starts from an unassigned node and traces forward along the dependencies. For the meat blanching process, its predecessor process is ingredient preparation; therefore, the predecessor tracing path is: meat blanching ← ingredient preparation. For the seasoning / reducing process, its predecessor process is stewing; stewing's predecessor is frying and browning; frying and browning's predecessor is meat blanching; and meat blanching's predecessor is ingredient preparation. Therefore, the predecessor tracing path is: seasoning / reducing ← stewing ← frying and browning ← meat blanching ← ingredient preparation.

[0062] The resource release propagation depth is determined based on the allocation status of each process node. Resource release propagation depth refers to the path length between an unassigned process node and an assigned process node that may release the required resources. For the meat blanching process, the required induction cooker is currently occupied by the frying and browning process; therefore, it needs to be traced back to the frying and browning process, resulting in a resource release propagation depth of 1. Similarly, the seasoning and sauce reduction process also requires an induction cooker, but it is occupied by the frying and browning process, so the resource release propagation depth is also 1. The allocation target equipment corresponding to the assigned process node with the largest resource release propagation depth is selected as the priority candidate equipment for the unassigned process node. In this embodiment, the priority candidate equipment for both the meat blanching and seasoning and sauce reduction processes is an induction cooker.

[0063] For unassigned process nodes with multiple predecessor tracing paths, extract the priority candidate equipment and corresponding resource release time points from each predecessor tracing path. In complex task execution graphs, unassigned processes may have multiple predecessor tracing paths. For example, if the seasoning and reducing process has another predecessor path: seasoning and reducing ← frying and browning, then these two paths need to be considered separately. The resource release time point refers to the moment after the priority candidate equipment completes the currently assigned process. For an induction cooker, the time to complete the frying and browning process is 25 minutes later.

[0064] The timing match degree between the path duration and resource release time of each precursor tracing path is calculated, and path priority is determined based on the timing match degree. Path duration refers to the total time required to execute along the path from the starting process to the unassigned process. The timing match degree is calculated as the closeness between the resource release time and the earliest executable time of the unassigned process; the closer, the higher the match degree. For the meat blanching process, its precursor process, ingredient preparation, requires 15 minutes, plus 2 minutes of preparation time between processes, so the earliest executable time for the meat blanching process is 17 minutes later. The induction cooker releases after 25 minutes, with a timing difference of 8 minutes. For the seasoning and sauce reduction process, calculated along the path of seasoning and sauce reduction ← stewing ← frying and coloring ← meat blanching ← ingredient preparation, the total time before seasoning and sauce reduction can be executed is 103 minutes: 15 minutes for ingredient preparation + 10 minutes for meat blanching + 25 minutes for frying and coloring + 45 minutes for stewing + 8 minutes for preparation time in each step. The induction cooker can release after 25 minutes, with a timing difference of 78 minutes. If there is a second path, namely seasoning and reducing sauce ← frying and coloring, the path duration is 25 minutes for frying and coloring + 2 minutes for preparation time = 27 minutes. The timing difference between this path and the resource release time of 25 minutes is 2 minutes, which is significantly less than the timing difference of the first path.

[0065] The priority candidate device corresponding to the predecessor tracing path with the highest path priority is selected as the optimal candidate device for the unassigned process node. Path priority is determined based on timing matching degree; paths with higher timing matching degree have higher priority. In this embodiment, the timing matching degree of the second path in the seasoning and reducing process is higher than that of the first path. Therefore, the induction cooker, the priority candidate device corresponding to the second path, is selected as the optimal candidate device for the seasoning and reducing process.

[0066] A delayed binding mapping relationship is established between the currently unassigned process nodes, priority candidate equipment, and their corresponding resource release time points to obtain resource release dependency information. The delayed binding mapping relationship indicates that the unassigned process will obtain the right to use a specific equipment at a specific time point. For the meat blanching process, a delayed binding mapping relationship is established: Meat blanching - Induction cooker - 25 minutes later. For the seasoning and sauce reduction process, a delayed binding mapping relationship is established: Seasoning and sauce reduction - Induction cooker - 25 minutes later. Since both processes are mapped to the same resource release time point, further decisions are needed based on process priority. The key quantification index for meat blanching is 0.71, and for seasoning and sauce reduction it is 0.65. Therefore, meat blanching prioritizes using the induction cooker, while seasoning and sauce reduction must wait until meat blanching is completed before using the induction cooker, i.e., seasoning and sauce reduction - induction cooker - (25+10+2)=37 minutes later. The final resource release dependency information is: meat blanching depends on the induction cooker being released during the frying and browning process, and seasoning and sauce reduction depends on the induction cooker being released during meat blanching.

[0067] In this embodiment, by arranging process nodes in descending order according to key quantifiable indicators, the scheduling process can prioritize critical processes that have a greater impact on the overall task, avoiding the problem of delayed allocation of critical processes leading to overall task chain delays. This improves overall execution efficiency and time controllability. By sequentially extracting the set of executable equipment from the feasibility mapping relationship and matching the remaining equipment capacity with the process resource requirements, the rationality of equipment allocation is effectively improved. By constructing the predecessor tracing path of unallocated process nodes and determining priority replacement equipment based on the resource release propagation depth of each process, potential shareable resources can be dynamically discovered by utilizing the resource release information of existing equipment allocation in the task chain. This solves the problem of being unable to handle temporary resource shortages or the difficulty in utilizing resource gaps in upstream and downstream processes, significantly improving resource reuse efficiency.

[0068] In one alternative implementation, Combinations of process nodes that do not have timing constraints with unassigned process nodes are identified to obtain parallel-executable process combinations. Based on the feasibility mapping relationship, kitchen appliances are then reassigned to these parallel-executable process combinations, generating a kitchen appliance scheduling sequence including: Based on the resource release dependency information, an adjacency matrix of the task execution graph is constructed and a transitive closure matrix is ​​calculated. Row vectors and column vectors corresponding to unassigned process nodes are extracted from the transitive closure matrix. The position indices of process nodes whose element values ​​in both row and column vectors are zero are used as candidate parallel nodes. Process nodes with the same priority backup equipment and a resource release time difference less than a preset time threshold are selected from the candidate parallel nodes. The resource feasibility of the process nodes and the remaining capacity parameters of the priority backup equipment are verified by combining the feasibility mapping relationship. The process nodes that pass the verification are combined to obtain a combination of parallel executable processes. Calculate the time dispersion of resource release time points corresponding to each process node in the parallelizable process combination. Calculate resource sufficiency based on the remaining capacity parameters of priority candidate devices and the combined resource requirements corresponding to the parallelizable process combination. Calculate the temporal resource synergy based on the time dispersion and resource sufficiency. Select the priority candidate device with the highest temporal resource synergy as the target parallel device. Allocate execution time windows based on the execution duration of the process nodes and the resource release time points of the target parallel devices. Generate a kitchen appliance scheduling sequence by topologically sorting each process node, its corresponding target device, and the execution time window according to the temporal dependency relationship.

[0069] Based on resource release dependency information, an adjacency matrix is ​​constructed from the task execution graph, and the transitive closure matrix is ​​calculated. The adjacency matrix is ​​a two-dimensional array representing the dependencies between process nodes. If process j depends on process i, the element in the i-th row and j-th column of the matrix is ​​1; otherwise, it is 0. Taking the braised pork ribs cooking task as an example, it includes five processes: ingredient preparation, blanching, frying and browning, stewing, and seasoning and reducing the sauce. According to the correct process dependencies, ingredient preparation is the first step; blanching depends on ingredient preparation; frying and browning depends on blanching; stewing depends on frying and browning; and seasoning and reducing the sauce depends on stewing. Combining the resource release dependency information, blanching needs to be performed after frying and browning releases the induction cooker, and seasoning and reducing the sauce needs to be performed after blanching is complete. Note that there is a conflict between process logic dependencies and resource release dependencies: blanching should logically precede frying and browning, but due to resource constraints, it must wait for frying and browning to release the induction cooker.

[0070] To resolve this conflict, the scheduling strategy needs to be modified. Correct resource allocation should respect the logical dependencies between processes, i.e., blanching the meat should be performed first, followed by pan-frying and browning. The revised initial scheduling allocation is as follows: Ingredient preparation - food processor; meat blanching - induction cooker; pan-frying and browning requires waiting for the meat blanching process to finish before releasing the induction cooker; stewing - electric pressure cooker; seasoning and reducing sauce requires waiting for the pan-frying process to finish before releasing the induction cooker. The resource release dependency information is corrected to: pan-frying and browning depends on the meat blanching process finishing before releasing the induction cooker; seasoning and reducing sauce depends on stewing being completed and pan-frying and browning being finished before releasing the induction cooker.

[0071] A 5×5 adjacency matrix A is constructed based on the corrected resource release dependency information. A[0][1]=1 indicates that blanching meat depends on ingredient preparation, A[1][2]=1 indicates that pan-frying and browning depends on blanching meat, A[2][3]=1 indicates that stewing depends on pan-frying and browning, and A[3][4]=1 indicates that seasoning and reducing sauce depends on stewing. These reflect the logical dependencies of the processes. The resource release dependency is expressed as A[1][2]=1 indicating that pan-frying and browning requires waiting for blanching meat to release the induction cooker, and A[2][4]=1 indicating that seasoning and reducing sauce requires waiting for pan-frying and browning to release the induction cooker.

[0072] The transitive closure matrix is ​​calculated to identify direct and indirect dependencies between processes. The Warshall algorithm is used, iterating through the adjacency matrix for five rounds to obtain the transitive closure matrix T. In T, T[i][j]=1 indicates that process j directly or indirectly depends on process i. Analysis of T reveals that frying and coloring, and seasoning and reducing are incompletely assigned process nodes, requiring the extraction of their corresponding row and column vectors.

[0073] Extract the row and column vectors corresponding to the unassigned process nodes from the transitive closure matrix. The row vector T[2]=[0,0,0,1,1] corresponding to frying and browning indicates that stewing and seasoning / reducing depend on frying and browning; the column vector [T[0][2],T[1][2],T[2][2],T[3][2],T[4][2]]=[1,1,0,0,0] indicates that frying and browning depends on ingredient preparation and blanching of meat. The row vector T[4]=[0,0,0,0,0] corresponding to seasoning / reducing indicates that no process depends on seasoning / reducing; the column vector [T[0][4],T[1][4],T[2][4],T[3][4],T[4][4]]=[1,1,1,1,0] indicates that seasoning / reducing depends on all other processes.

[0074] The position indices of process nodes where all elements in the row and column vectors are zero are selected as candidate parallel nodes. For frying and browning, the elements at positions 0 and 1 in the row vector are zero, and the elements at positions 3 and 4 in the column vector are zero. Taking the intersection yields no result, indicating that frying and browning cannot currently be parallelized with other processes. For seasoning and reducing sauce, all elements in the row vector are zero, and the element at position 4 in the column vector is zero. This only involves seasoning and reducing sauce itself, indicating that under the existing dependencies, there are no other processes that can be executed in parallel with these two processes.

[0075] Re-analyze potential opportunities for parallel execution. Note that there is no direct dependency between ingredient preparation and the post-blanching meat process, and different equipment is used. Ingredient preparation uses a food processor, while the post-blanching meat process uses an induction cooker or electric pressure cooker. Therefore, some steps in ingredient preparation and the post-blanching meat process can be performed in parallel. Similarly, stewing uses an electric pressure cooker, while searing, browning, and seasoning / reducing sauces use an induction cooker; these steps also present the possibility of parallel execution.

[0076] From the candidate parallel nodes, priority is selected for process nodes with the same alternative equipment and a resource release time difference less than a preset time threshold. In this example, stewing and frying / browning use different equipment and can be executed in parallel; ingredient preparation and meat blanching use different equipment and can also be executed in parallel. Assuming a preset time threshold of 15 minutes, the execution time of each process is analyzed: ingredient preparation requires 15 minutes, meat blanching requires 10 minutes, frying / browning requires 25 minutes, stewing requires 45 minutes, and seasoning / reducing requires 15 minutes. Based on process logic dependencies and equipment resource constraints, the following parallel execution combinations can be determined: ingredient preparation (food processor) and meat blanching (induction cooker) can be executed in parallel; frying / browning (induction cooker) and stewing (electric pressure cooker) can be executed in parallel.

[0077] Resource feasibility was verified by combining the resource requirement parameters of each process node with the remaining capacity parameters of the equipment, based on the feasibility mapping relationship. Food preparation requires a food processor with a power of 500 watts, and blanching meat requires an induction cooker with a temperature of 100℃ and a power of 800 watts; the food processor has a remaining power of 600 watts, and the induction cooker has a remaining temperature capacity of 250℃ and a power of 1000 watts, so the verification passed. Frying and browning requires an induction cooker with a temperature of 170℃ and a power of 1000 watts, and stewing requires an electric pressure cooker with a temperature of 100℃ and a power of 800 watts; the remaining capacity of the induction cooker and electric pressure cooker meets the requirements, so the verification passed.

[0078] Calculate the time dispersion of resource release times for each process node in a parallelizable process combination. For the ingredient preparation and meat blanching combination, both processes start execution at 0 minutes. Ingredient preparation releases the food processor at 15 minutes, and meat blanching releases the induction cooker at 10 minutes, with a time dispersion of 6.25. For the frying and browning and stewing combination, frying and browning executes from 10 minutes to 35 minutes, and stewing executes from 10 minutes to 55 minutes, with a time dispersion of 100. The smaller the time dispersion, the more concentrated the process execution time, and the greater the parallelization potential.

[0079] Resource sufficiency is calculated based on the remaining capacity parameters of the equipment and the combined resource requirements corresponding to the combinations of processes that can be executed in parallel. For the combination of food preparation and meat blanching, the food processor has a remaining power capacity of 600 watts and a power requirement of 500 watts, resulting in a sufficiency of 1.2; the induction cooker has a remaining temperature capacity of 250℃ and a temperature requirement of 100℃, resulting in a sufficiency of 2.5. For the combination of frying and stewing, the induction cooker has a power sufficiency of 1 and a temperature sufficiency of 1.47; the electric pressure cooker has a power sufficiency of 1.13 and a temperature sufficiency of 1.2. A resource sufficiency greater than 1 indicates that resources are sufficient to support execution.

[0080] The temporal resource synergy degree is calculated based on time dispersion and resource sufficiency. This degree comprehensively evaluates the time and resource conditions for parallel execution of processes and is calculated as a weighted ratio of resource sufficiency to time dispersion; a higher value indicates higher synergy. For the combination of ingredient preparation and meat blanching, the calculated temporal resource synergy degree is 0.32; for the combination of frying and stewing, the calculated degree is 0.012. The combination with the highest temporal resource synergy degree is selected for parallel scheduling, meaning ingredient preparation and meat blanching are executed in parallel.

[0081] The execution time window is allocated based on the execution duration of the process nodes and the resource release time of the equipment. The execution time window defines the start and end times of the process. The revised scheduling scheme is as follows: Ingredient preparation - food processor - [0, 15] minutes, meat blanching - induction cooker - [0, 10] minutes, pan-frying and browning - induction cooker - [10, 35] minutes, stewing - electric pressure cooker - [10, 55] minutes (in parallel with pan-frying and browning), seasoning and reducing sauce - induction cooker - [55, 70] minutes. This scheme respects the logical dependencies between processes and achieves efficient utilization of equipment resources.

[0082] The kitchen appliance scheduling sequence is generated by topologically sorting each process node, its corresponding target equipment, and execution time window according to their temporal dependencies. Topological sorting ensures that dependencies are satisfied, meaning that a process can only begin execution after all its predecessor processes have been completed.

[0083] In this embodiment, by constructing the adjacency matrix of the task execution graph and calculating the transitive closure, the reachability relationships between process nodes can be systematically identified. Parallelizable processes on different branches can be globally and automatically identified, improving the completeness and accuracy of parallel scheduling identification. Candidate parallel nodes are screened by combining resource release dependency information, and the matching relationship between resource requirement parameters and remaining equipment capacity is further verified. This ensures that the selection of parallel combinations not only satisfies the absence of process dependencies but also guarantees resource-level executability. By calculating the time dispersion and resource sufficiency of parallelizable process combinations and constructing temporal resource synergy based on these two factors, the advantages and disadvantages of parallel scheduling schemes can be comprehensively evaluated from both time synchronization and resource utilization efficiency perspectives. By selecting the equipment with the highest temporal resource synergy as the parallel execution target and allocating specific execution time windows based on resource release time points, the idle time periods of equipment can be utilized to the maximum extent, achieving efficient overlapping use of equipment resources and improving the overall task execution parallelism and timeliness.

[0084] Figure 2 This is a flowchart illustrating the time-series resource collaborative scheduling process of the intelligent scheduling and control method for collaborative work of kitchen appliance groups according to an embodiment of the present invention.

[0085] A second aspect of the present invention provides an intelligent scheduling and control system for collaborative operation of kitchen appliances, comprising: The first unit is used to acquire the operating status information and cooking task information of each kitchen appliance in the kitchen appliance group, and to decompose the cooking task information to be executed into processes in combination with preset cooking process constraint rules, identify the temporal dependency relationship between different processes to obtain a task execution diagram, extract the capability parameters of the kitchen appliances based on the operating status information and match them with the resource requirement parameters of the process nodes in the task execution diagram to obtain a feasibility mapping relationship. The second unit is used to traverse backward along the temporal dependency path in the task execution graph starting from each process node, count the number of subsequent processes corresponding to the current process node, accumulate the longest time span from the current process node to the end of the task execution graph, and calculate key metric indicators based on the number of subsequent processes and the longest time span. The third unit is used to allocate kitchen appliances to the process nodes according to the feasibility mapping relationship and the key quantification index to obtain a preliminary scheduling allocation result, extract the unallocated process nodes in the preliminary scheduling allocation result and backtrack along the temporal dependency relationship in the task execution graph to obtain resource release dependency information, identify process node combinations that do not have temporal order constraints with the unallocated process nodes to obtain parallel executable process combinations, and re-allocate kitchen appliances to the parallel executable process combinations in combination with the feasibility mapping relationship to generate a kitchen appliance scheduling sequence.

[0086] A third aspect of the present invention provides an electronic device, comprising: A processor and a memory for storing processor-executable instructions, wherein the processor is configured to invoke instructions stored in the memory to perform the aforementioned method.

[0087] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0088] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent scheduling and control of collaborative work among kitchen appliances, characterized in that: include: The system acquires the operating status information and cooking task information of each kitchen appliance in the kitchen appliance group. It then decomposes the cooking task information into processes based on preset cooking process constraints, identifies the temporal dependencies between different processes to obtain a task execution graph, extracts the capability parameters of the kitchen appliances based on the operating status information, and matches and judges them with the resource requirement parameters of the process nodes in the task execution graph to obtain a feasibility mapping relationship. Starting from each process node, traverse backward along the temporal dependency path in the task execution graph, count the number of subsequent processes corresponding to the current process node, accumulate the longest time span from the current process node to the end of the task execution graph, and calculate the key metric based on the number of subsequent processes and the longest time span. Based on the feasibility mapping relationship and the key quantification indicators, kitchen appliances are allocated to the process nodes to obtain preliminary scheduling and allocation results. Unallocated process nodes in the preliminary scheduling and allocation results are extracted and resource release dependency information is obtained by backtracking along the temporal dependency relationship in the task execution graph. Process node combinations that do not have temporal sequence constraints with unallocated process nodes are identified to obtain parallel execution process combinations. Kitchen appliances are then allocated again to parallel execution process combinations based on the feasibility mapping relationship to generate a kitchen appliance scheduling sequence.

2. The method according to claim 1, characterized in that, The system acquires the operating status information and cooking task information of each kitchen appliance in the kitchen appliance group, and decomposes the cooking task information into processes based on preset cooking process constraints. It then identifies the temporal dependencies between different processes to obtain a task execution graph, including: The system receives a cooking task input by the user and determines the dish identifier and cooking target parameters in the cooking task. Based on the dish identifier, it queries the preset cooking process constraint rules and extracts the process sequence corresponding to the dish identifier. The process sequence is then decomposed and the operation type and resource requirement parameters corresponding to each process are identified. Based on the process flow definition in the cooking process constraint rules, the preceding and subsequent processes of each process in the process sequence are analyzed. Based on the preceding and subsequent processes, the state transmission relationship between different processes is determined and a preliminary dependency graph is constructed. Obtain the operating status information of each kitchen appliance in the kitchen appliance group, extract the equipment occupancy time and remaining processing capacity from the operating status information, traverse the process nodes in the preliminary dependency graph and extract the resource requirement parameters corresponding to each process node, compare the resource requirement parameters with the remaining processing capacity corresponding to each kitchen appliance, identify the process whose resource requirement parameters exceed the remaining processing capacity, and determine the temporal dependency relationship between different processes to obtain the task execution graph.

3. The method according to claim 1, characterized in that, Based on the operational status information, the capability parameters of the kitchen appliances are extracted and matched with the resource requirement parameters of the process nodes in the task execution diagram to obtain the feasibility mapping relationship, including: The system analyzes the operating status information to extract the equipment type identifier and historical execution record of the kitchen appliances and determines the capability parameters of each kitchen appliance. The capability parameters include a set of functional operations and performance boundary values. The system traverses the process nodes in the task execution graph and extracts the operation type and resource requirement parameters of each process node. The system matches the operation type with the set of functional operations and identifies candidate kitchen appliances that contain the operation type of the process node in the set of functional operations. Retrieve historical process nodes that the candidate kitchen appliance has performed with the same operation type as the current process node from the historical execution records, extract the historical resource requirement parameters and actual resource consumption corresponding to the historical process nodes, and perform regression fitting to obtain a regression mapping relationship. Substitute the resource requirement parameters of the current process node into the regression mapping relationship to calculate the predicted resource consumption. The feasibility of the pre-acquired preceding and subsequent processes on the candidate kitchen appliances is determined, and the total number of executable nodes is counted. A topology consistency score is calculated based on the total number of nodes. A resource satisfaction degree is calculated based on the predicted resource consumption and the performance boundary value, and a weighted sum is obtained with the topology consistency score to obtain a matching adaptation value. Candidate kitchen appliances with matching adaptation values ​​higher than a preset matching threshold are associated with process nodes to obtain a feasibility mapping relationship.

4. The method according to claim 1, characterized in that, Starting from each process node, traverse backwards along the temporal dependency path in the task execution graph, count the number of subsequent processes corresponding to the current process node, and accumulate the longest time span from the current process node to the end of the task execution graph. Based on the number of subsequent processes and the longest time span, calculate the key metrics, including: Starting from each process node, traverse backward along the time-dependent path in the task execution graph, extract the estimated execution time of each subsequent process node until the endpoint node is reached, accumulate the estimated execution time of each process node in each reachable path to obtain the path duration and determine the maximum time span, and count the number of all subsequent process nodes reachable from the current process node to obtain the number of subsequent processes. Extract a set of historical process nodes with the same operation type as the current process node from the pre-acquired historical execution records. Obtain the temporal extension depth and corresponding task completion deviation value of each historical process node and quantify the correlation to obtain the time dimension sensitive factor. Obtain the topological coverage breadth of each historical process node and corresponding task completion deviation value and quantify the correlation to obtain the breadth dimension sensitive factor. A time-series criticality metric is obtained by nonlinearly coupling the longest time span with a time-dimensional sensitivity factor, and a topological impact metric is obtained by nonlinearly coupling the number of subsequent processes with a breadth-dimensional sensitivity factor. A basic priority score is calculated based on the time-series criticality metric and the topological impact metric. The basic priority score is encoded as a particle position vector. The criticality metric is obtained by iteratively optimizing the particle position vector using a particle swarm optimization algorithm, combined with the resource requirement parameters corresponding to the current process node and the remaining capacity parameters corresponding to the kitchen appliances.

5. The method according to claim 4, characterized in that, The basic priority score is encoded as a particle position vector. Combined with the resource requirement parameters corresponding to the current process node and the remaining capacity parameters corresponding to the kitchen appliances, the key metric indicators are obtained by iteratively optimizing the particle position vector using a particle swarm optimization algorithm. These indicators include: The basic priority scores of each process node are vectorized and encoded according to the arrangement order of the process nodes in the task execution diagram to obtain the initial position vector. The constraint penalty term is determined based on the matching tension between the resource demand parameter and the remaining capacity parameter. With the goal of minimizing the task completion time, the fitness corresponding to the initial position vector is calculated as the individual optimal fitness in combination with the pre-set constraint penalty term. The particle position vector with the largest individual optimal fitness is selected as the global optimal vector. The individual cognitive component is determined based on the deviation between the particle's current position vector and the individual optimal vector, as well as a preset individual learning factor. The social cognitive component is determined based on the deviation between the particle's current position vector and the global optimal vector, as well as a preset global influence coefficient. The particle velocity is updated based on the individual cognitive component and the social cognitive component to obtain an updated velocity vector. The candidate position vector is obtained by fusing the particle's current position vector and the updated velocity vector. The candidate position vectors are tested for feasibility region. Based on the test results, new position vectors are determined. The fitness of the new position vectors is calculated and compared with the individual optimal fitness. Based on the comparison results, the individual optimal vectors and the global optimal vectors are updated. The updates are repeated until the preset maximum number of iterations is reached. The priority values ​​of each process node in the global optimal vector are extracted to obtain key metric indicators.

6. The method according to claim 1, characterized in that, Based on the feasibility mapping relationship and the key quantification indicators, kitchen appliances are allocated to the process nodes to obtain preliminary scheduling and allocation results. Unallocated process nodes in the preliminary scheduling and allocation results are extracted, and resource release dependency information is obtained by backtracking along the temporal dependency relationship in the task execution graph, including: Each process node is sorted in descending order according to key quantifiable indicators. Based on the sorting results, the set of executable kitchen appliances corresponding to each process node is extracted from the feasibility mapping relationship. Based on the matching relationship between the remaining capacity parameter and the resource demand parameter, kitchen appliances are allocated to the process nodes to obtain the preliminary scheduling and allocation results. Extract all unassigned process nodes from the preliminary scheduling and allocation results, construct a predecessor tracing path by backtracking along the temporal dependency in the task execution graph, determine the resource release propagation depth based on the allocation status of each process node, and select the allocation target device corresponding to the allocated process node with the largest resource release propagation depth as the priority candidate device for the unassigned process node. For an unassigned process node with multiple predecessor tracing paths, the priority candidate device and the corresponding resource release time point in each predecessor tracing path are extracted. The timing matching degree between the path duration and the resource release time point of each predecessor tracing path is calculated, and the path priority is determined based on the timing matching degree. The priority candidate device corresponding to the predecessor tracing path with the highest path priority is taken as the optimal candidate device for the unassigned process node. A delayed binding mapping relationship is established between the current unassigned process node and the priority candidate device and the corresponding resource release time point to obtain resource release dependency information.

7. The method according to claim 1, characterized in that, Combinations of process nodes that do not have timing constraints with unassigned process nodes are identified to obtain parallel-executable process combinations. Based on the feasibility mapping relationship, kitchen appliances are then reassigned to these parallel-executable process combinations, generating a kitchen appliance scheduling sequence including: Based on the resource release dependency information, an adjacency matrix of the task execution graph is constructed and a transitive closure matrix is ​​calculated. Row vectors and column vectors corresponding to unassigned process nodes are extracted from the transitive closure matrix. The position indices of process nodes whose element values ​​in both row and column vectors are zero are used as candidate parallel nodes. Process nodes with the same priority backup equipment and a resource release time difference less than a preset time threshold are selected from the candidate parallel nodes. The resource feasibility of the process nodes and the remaining capacity parameters of the priority backup equipment are verified by combining the feasibility mapping relationship. The process nodes that pass the verification are combined to obtain a combination of parallel executable processes. Calculate the time dispersion of resource release time points corresponding to each process node in the parallelizable process combination. Calculate resource sufficiency based on the remaining capacity parameters of priority candidate devices and the combined resource requirements corresponding to the parallelizable process combination. Calculate the temporal resource synergy based on the time dispersion and resource sufficiency. Select the priority candidate device with the highest temporal resource synergy as the target parallel device. Allocate execution time windows based on the execution duration of the process nodes and the resource release time points of the target parallel devices. Generate a kitchen appliance scheduling sequence by topologically sorting each process node, its corresponding target device, and the execution time window according to the temporal dependency relationship.

8. An intelligent scheduling and control system for collaborative operation of kitchen appliances, used to implement the method of any one of claims 1-7, characterized in that, include: The first unit is used to acquire the operating status information and cooking task information of each kitchen appliance in the kitchen appliance group, and to decompose the cooking task information to be executed into processes in combination with preset cooking process constraint rules, identify the temporal dependency relationship between different processes to obtain a task execution diagram, extract the capability parameters of the kitchen appliances based on the operating status information and match them with the resource requirement parameters of the process nodes in the task execution diagram to obtain a feasibility mapping relationship. The second unit is used to traverse backward along the temporal dependency path in the task execution graph starting from each process node, count the number of subsequent processes corresponding to the current process node, accumulate the longest time span from the current process node to the end of the task execution graph, and calculate key metric indicators based on the number of subsequent processes and the longest time span. The third unit is used to allocate kitchen appliances to the process nodes according to the feasibility mapping relationship and the key quantification index to obtain a preliminary scheduling allocation result, extract the unallocated process nodes in the preliminary scheduling allocation result and backtrack along the temporal dependency relationship in the task execution graph to obtain resource release dependency information, identify process node combinations that do not have temporal order constraints with the unallocated process nodes to obtain parallel executable process combinations, and re-allocate kitchen appliances to the parallel executable process combinations in combination with the feasibility mapping relationship to generate a kitchen appliance scheduling sequence.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images