Concrete mixing plant production management system

Abstract

The invention discloses a concrete mixing plant production management system comprising the following modules: a discrete event modeling module used for constructing a discrete event simulation model; the data acquisition module is used for acquiring a real-time operation data set of the concrete mixing plant and inputting the real-time operation data set as a discrete event simulation model; the simulation execution module is used for executing an event-driven simulation process; the ant colony optimization module is used for constructing a scheduling optimization objective function and a path heuristic function according to the simulation data set, and executing search to generate an optimal material sending scheduling sequence; the scheduling execution module is used for executing a scheduling process according to the optimal material sending scheduling sequence; and the feedback control module is used for monitoring the execution deviation, transmitting the latest operation state back to the discrete event modeling module, triggering scheduling reconstruction, and realizing the production dynamic closed-loop optimization management of the concrete mixing plant. According to the invention, continuity and dynamic optimality of production scheduling are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of mixing plants, and particularly to a production management system for a concrete mixing plant. Background Art

[0002] With the acceleration of infrastructure construction and urbanization, the role of concrete mixing plants in the construction industry has become increasingly prominent. The efficiency of its production scheduling and material feeding management is directly related to the project progress and resource utilization level. To improve the scheduling efficiency, some enterprises have tried to introduce a concrete mixing plant management system based on an information platform. By deploying basic scheduling modules and data acquisition devices, process control is implemented for key processes such as mixing, loading, transportation, and unloading. However, the existing production management systems for concrete mixing plants generally rely on static rule setting and strategy models driven by manual experience, and it is difficult to cope with the dynamically changing production environment.

[0003] On the one hand, most current systems use a fixed priority strategy to plan the process scheduling, and fail to establish a dynamic scheduling mechanism for real-time perception and response to the system operation state, resulting in queuing congestion, resource conflicts, and feeding delays often occurring during the peak material feeding period. On the other hand, although some systems integrate real-time data acquisition modules, the collected data is not systematically used for feedback control and scheduling optimization, resulting in the scheduling results lacking adaptability to on-site fluctuations, the system lacking closed-loop adjustment ability, and being unable to automatically adjust the scheduling strategy according to bottleneck process changes or resource allocation imbalances.

[0004] In addition, the existing system scheduling optimization methods generally do not introduce advanced intelligent algorithms for path search and multi-objective trade-off. The scheduling results often do not have global optimality, and it is also difficult to ensure the full utilization of equipment resources and task concurrency coordination ability under high load conditions. The system has significant performance bottlenecks when facing complex task scenarios.

[0005] In summary, there is an urgent need for a comprehensive management method that can combine real-time data perception, dynamic simulation analysis, and intelligent scheduling optimization to improve the overall operation efficiency and response ability of the mixing plant. Summary of the Invention

[0006] An object of the present invention is to provide a production management system for a concrete mixing plant, and the present invention realizes the continuity and dynamic optimality of production scheduling.

[0007] A production management system for a concrete mixing plant according to an embodiment of the present invention includes modules: A discrete event modeling module for constructing a discrete event simulation model, where the discrete event simulation model includes a set of key processes, a set of resources, a set of process execution dependencies, a set of process event data, and a set of process trigger functions; A data acquisition module, which is used to obtain the real-time operation data set of the concrete mixing plant and use the real-time operation data set as the input of the discrete event simulation model; A simulation execution module, which is used to execute an event-driven simulation process and generate a simulation data set including process queue length functions, resource utilization functions, and bottleneck functions; An ant colony optimization module, which is used to construct a scheduling optimization objective function and a path heuristic function based on the simulation data set, and execute a search to generate an optimal material feeding scheduling sequence; A scheduling execution module, which is used to execute the scheduling process according to the optimal material feeding scheduling sequence and control equipment feeding according to the process sequence, resource allocation, and time constraints; A feedback control module, which is used to monitor the execution deviation, transmit the latest operating state back to the discrete event modeling module, trigger scheduling reconstruction, and realize the dynamic closed-loop optimization management of the production of the concrete mixing plant.

[0008] A production management method for a concrete mixing plant, which is applied to a production management system of a concrete mixing plant, and includes the following steps: S1. Construct a discrete event simulation model of the production process of the concrete mixing plant. The discrete event simulation model models the discrete events of the key processes of concrete mixing, loading, transportation, and unloading, and defines the triggering conditions, execution order, and resource constraints of each discrete event; S2. Collect the real-time operation data set of the concrete mixing plant. The real-time operation data set includes the operating status of each device, the arrival time of transport vehicles, the loading completion time, and the number of queuing vehicles, and use the collected real-time operation data set as the input data of the discrete event simulation model; S3. Input the real-time operation data set into the discrete event simulation model, execute the discrete event simulation, and generate simulation data reflecting the material feeding process of the concrete mixing plant. The simulation data covers queuing status, equipment utilization rate, and potential scheduling bottleneck information; S4. Initialize the ant colony optimization algorithm based on the simulation data, and at the same time define the scheduling optimization objective. Use the initialized ant colony optimization algorithm to perform scheduling solution based on the simulation data, and generate an optimal material feeding scheduling sequence through iterative search; S5. Apply the optimal material feeding scheduling sequence to the scheduling execution module of the concrete mixing plant to control the execution of the actual feeding process, so that the actual feeding scheduling is strictly carried out in the optimized order, and realize the management of the material feeding queue of the concrete mixing plant; S6. When it is detected that there is a significant deviation between the actual operation state and the simulation data, automatically feedback the newly collected real-time operation data set to the discrete event simulation model, update the running state of the discrete event simulation model, and restart the closed-loop scheduling optimization process from step S4 to step S5 to realize the dynamic real-time adjustment of the material feeding queue control of the concrete mixing plant.

[0009] Optionally, S1 includes the following steps: S11. Set the key process set of the concrete mixing plant. The key process set includes the mixing process, loading process, transportation process, and unloading process. The key process set serves as the basic process unit in the discrete event simulation model; S12. Define the discrete event simulation model as a structured modeling unit consisting of five components. The five components include the key process set, resource set, process execution dependency set, process event data set, and process trigger function set. The five components together constitute the discrete event simulation model for simulating the process behavior in the material delivery process of the concrete mixing plant; S13. Construct a process execution dependency graph. In the process execution dependency graph, the execution of each process depends on the completion of the previous process. If the execution of process A depends on the completion of process B, then record the dependency relationship pointing from process B to process A in the process execution dependency set. The process execution dependency graph is used to constrain the timing sequence of process triggering; S14. Set the resource set of the concrete mixing plant. The resource set includes various resources such as mixing equipment, loading systems, transport vehicles, and unloading platforms. Each type of resource has a unique resource identifier and a maximum available capacity. The maximum available capacity is used to determine whether the resources required during process execution meet the allocation conditions; S15. Construct a process event data set. Each data record in the process event data set includes the process type, process arrival time, process start time, process end time, and the resource identifier occupied. The process event data set is used to drive the process state change process in the discrete event simulation model; S16. Define a process trigger function set for each process. The process trigger function set is used to determine whether the current system simulation time reaches the moment when the process is allowed to be triggered, and to determine whether the required resources have available capacity. If both the time condition and the resource condition are met, the output of the process trigger function is triggerable, otherwise it is non-triggerable; S17. At any system simulation moment, construct the current triggerable process queue according to the results of the process trigger function. The triggerable process queue contains all the processes that meet the trigger conditions at the current moment. Define a process priority function. The process priority function is calculated based on the cumulative waiting time of the process, the idle time of the required resources, and the process urgency. Sort the processes in the triggerable process queue according to the results of the process priority function, and preferentially schedule the processes with higher rankings to achieve coordinated scheduling of resource conflicts; S18. Set a concurrent execution time window for each process. The concurrent execution time window defines the earliest start time and the latest completion time that a process is allowed in the system simulation. In each system simulation cycle, if the time windows of two or more processes overlap, the required resources have no conflicts, and the sequence constraints in the process execution dependency graph are not violated, then such processes are classified into the concurrent process set, and the processes in the concurrent process set are allowed to advance and execute simultaneously.

[0010] Optionally, the S3 includes the following steps: S31. Input the real-time operation data set into the discrete event simulation model, perform corresponding structure mapping on the real-time operation data set and the process event data set and update the process status according to the current simulation time of the system, which is used to drive the process trigger judgment; S32. At the current simulation time of the system, perform an event-driven time advancement simulation operation. Relying on the key process set, resource set, process execution dependency set, process event data set, and process trigger function set defined in the discrete event simulation model, collect all processes in the waiting state and not meeting the trigger conditions to form a queued process set; S33. Based on the queued process set, perform statistics on each type of process to calculate the queued quantity of this type of process at the current simulation time, and obtain the process queuing length function . The process queuing length function is classified by process type and calculates the number of processes of this type that have not met the trigger conditions at the current system simulation time, which is used to represent the waiting load status of different process types; S34. Analyze the status of each type of resource in the resource set at the current simulation time to calculate the quantity occupied by the executed processes of this type of resource, and combine it with the maximum available capacity of this type of resource to obtain the resource utilization rate function . The resource utilization rate function is composed of the ratio between the current occupied quantity of the resource and the maximum available capacity of the resource, which is used to measure the current operating load level of this type of resource; S35. Construct a process bottleneck function . The process bottleneck function takes the process queuing length function and the resource utilization rate function of the corresponding resource type at the current simulation time as inputs and is formed according to the ratio relationship between the two, which is used to reflect the queuing accumulation degree of this process when the resource is in a low utilization state. The larger the value of the process bottleneck function, the more serious the current restriction of this process on the system scheduling; S36. Within the total time period of the system simulation, perform cumulative average calculation on the time series of the process bottleneck function to form the process bottleneck index ; S37. Output the simulation data set , where the simulation data set includes the system simulation time series, the process queue length function for each type of process, the resource utilization rate function for each type of resource, the process bottleneck function for each type of process, and the process bottleneck index. The simulation data set is used to reflect the operating status of each key process and resource in the concrete mixing plant's material distribution process driven by the current real-time operation data set.

[0011] Optionally, the process bottleneck index is based on the numerical sequence of the process bottleneck function throughout the simulation period. By averaging the bottleneck function values at all time nodes, a long-term evaluation index is obtained to reflect the degree of constraint of the process on the system operation in the entire material distribution process.

[0012] Optionally, S4 includes the following steps: S41. Initialize the parameter set of the ant colony optimization algorithm; S42. Construct the scheduling optimization objective function. The scheduling optimization objective function takes the process queue length function for each type of process, the resource utilization rate function for each type of resource, and the process bottleneck function for each type of process in the simulation data set as inputs. The structure of the scheduling optimization objective function is used to comprehensively minimize the queuing waiting time, improve the resource load balance, and reduce the process bottleneck degree, constituting a global evaluation index for the scheduling path; S43. Construct the scheduling state space based on the queuing process set output in the simulation data set. Each state in the scheduling state space corresponds to a legal arrangement of the process scheduling order, and the scheduling order arrangement must satisfy the process execution dependency constraint and the resource availability constraint; S44. Construct the path heuristic function. The path heuristic function consists of three parts, namely the process bottleneck index item, the process waiting time item, and the resource response time item. The bottleneck index item is set according to the bottleneck index value corresponding to the current process. The process waiting time item is set according to the cumulative waiting duration of the current process in the process queue length function. The resource response time item is set according to the idle response time calculated in the resource utilization rate function for the resource type required by the process. The path heuristic function is used to guide the ants to preferentially select processes with less system pressure during the construction of the scheduling path; S45. Initialize the ant colony search structure. Each ant randomly selects a process from the scheduling state space as the starting node and gradually constructs a complete process scheduling path according to the combined weight of the calculated value of the path heuristic function and the pheromone concentration. All ants generate a complete candidate solution in each round of iteration; S46. After each round of iteration, the paths of all ants are scored according to the scheduling optimization objective function constructed in step S42, and the pheromone concentration on the corresponding paths is updated according to the scoring results. The update method includes two processes: pheromone intensification and pheromone evaporation. Pheromone intensification is enhanced by positive feedback according to the value of the scheduling optimization objective function, and pheromone evaporation is used to limit local over-concentration and avoid falling into local optima; S47. When the ant colony optimization algorithm reaches the set maximum number of iterations or the optimal path remains unchanged for several consecutive rounds, the current optimal scheduling path is output as the optimal material delivery scheduling sequence generated under the current simulation state. The optimal material delivery scheduling sequence is used to guide the execution of the material delivery process in the concrete mixing plant.

[0013] Optionally, the ant colony optimization algorithm parameter set includes the number of ants, the initial pheromone value, the pheromone evaporation coefficient, the path selection control factor, and the maximum number of iterations. The number of ants is set according to the total number of queuing processes at the current simulation time, and the initial pheromone value is based on the process bottleneck index corresponding to each type of process Set, and the pheromone evaporation coefficient is set according to the average value of the overall resource utilization rate of the current system.

[0014] Optionally, S5 includes the following steps: S51. Input the optimal material delivery scheduling sequence into the concrete mixing plant scheduling execution module. The optimal material delivery scheduling sequence consists of multiple scheduling units, and each scheduling unit includes a scheduling process identifier, a scheduling start time, the required resource number, and a resource allocation strategy; S52. According to the arrangement order of each scheduling unit in the optimal material delivery scheduling sequence, construct a delivery execution queue and load it into the scheduling queue control unit. The scheduling execution module activates the corresponding scheduling instruction when the following three scheduling conditions are met: All previous processes of the current scheduling process have been completed, meeting the process execution dependency set defined in the discrete event simulation model; The available capacity of the resources required for the scheduling process at the current moment in the resource set meets the resource allocation strategy; The system simulation time reaches or exceeds the scheduling start time corresponding to the scheduling process; S53. When the three scheduling conditions are simultaneously met, the scheduling execution module sends a start instruction to the device corresponding to the specified resource number in the scheduling unit to control the actual process execution; if any condition is not met, the scheduling unit pauses execution and enters the waiting state; S54. During the operation of the material delivery process, the scheduling execution module continuously detects whether the system execution status is consistent with the scheduling order. When it is found that there is a deviation between the execution order and the scheduling order, the sending of subsequent scheduling instructions is suspended, and the system scheduling status is rolled back to the position of the most recent completed scheduling process, waiting for the system status to realign with the scheduling logic again; S55. Manage the process scheduling according to the optimized scheduling order of the control concrete mixing plant scheduling execution module.

[0015] Optionally, the S6 includes the following steps: S61. During the actual execution of the material delivery process in the concrete mixing plant, collect the current actual operation status data set for status comparison and analysis with the process queuing length function, resource utilization function, and process bottleneck function included in the simulation data set ; S62. Define the actual operation status data set , and compare the current actual operation status data set with the simulation data set process by process and resource by resource. If any of the following conditions is met, it is determined that the system status deviation is significant: The actual completion time of a certain process in the actual operation status data set is delayed by more than the set threshold of the predicted completion time in the simulation data; The continuous occupation duration of a certain resource in the actual operation status data set is greater than the upper limit of the corresponding resource occupation time in the simulation data; The queuing time continuously exceeds the sum of the historical mean and standard deviation of the queuing length function in the simulation state in multiple consecutive processes; S63. When it is judged that the system deviation is greater than the threshold, immediately trigger the scheduling feedback mechanism and write the current system status into the latest real-time operation data set ; S64. Input the latest real-time operation data set into the discrete event simulation model, reconstruct the simulation input state, execute the discrete event simulation process, generate a new simulation data set , and update the process queuing length function, resource utilization function, and process bottleneck function in the simulation data; S65. Based on the updated simulation data set , re-initialize the ant colony optimization, path search, and optimal scheduling path output process to obtain a new optimal material delivery scheduling sequence; S66. Use the new optimal material delivery scheduling sequence as the feedback scheduling instruction and re-enter it into the concrete mixing plant scheduling execution module to re-manage the process execution, replace the old scheduling path, and achieve the dynamic closed-loop optimization management of the concrete mixing plant production.

[0016] The beneficial effects of the present invention are as follows: The present invention introduces a five - tuple structure including "key process set, resource set, process execution dependency set, process event data set, and process trigger function set" into the modeling of a concrete mixing plant, realizing the granularity modeling of key processes such as mixing, loading, transportation, and unloading. It can accurately determine the triggerable process queue at each system simulation moment, and dynamically sort in combination with resource availability and process priority functions, realizing the full - process modeling and real - time calculation of process conflicts, concurrent scheduling, and bottleneck judgment. By constructing a simulation data set, it outputs simulation indicators such as process queuing length function, resource utilization rate function, and bottleneck function, providing reliable global dynamic data support for subsequent optimization.

[0017] The present invention designs a composite path heuristic function composed of process bottleneck index items, process waiting time items, and resource response time items, and dynamically adjusts the pheromone evaporation coefficient and reinforcement mechanism according to the process queuing function and resource utilization rate function, enhancing the algorithm's perception ability of scheduling bottleneck nodes and effectively avoiding falling into a locally optimal scheduling sequence.

[0018] The present invention establishes a complete set of scheduling deviation detection and feedback closed - loop optimization system. When monitoring state deviations such as "process completion delay", "abnormal continuous occupation of resources", or "continuous queuing of multiple processes exceeding the threshold", it can automatically trigger scheduling reconstruction. The system will remap the newly collected real - time operation data set to the simulation model, reconstruct the simulation state in real - time, and start the ant colony algorithm for path re - search, outputting a new optimal scheduling path to replace the old execution path, thus realizing the continuity and dynamic optimality of production scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings: Figure 1 is a flowchart of a production management system for a concrete mixing plant proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] Now, the present invention will be further described in detail with reference to the drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0021] Refer to Figure 1 , a production management system for a concrete mixing plant, including modules: A discrete - event modeling module for constructing a discrete - event simulation model, where the discrete - event simulation model includes a key process set, a resource set, a process execution dependency set, a process event data set, and a process trigger function set; A data acquisition module, which is used to obtain the real-time operation data set of the concrete mixing plant and use the real-time operation data set as the input of the discrete event simulation model; A simulation execution module, which is used to execute an event-driven simulation process and generate a simulation data set including the process queuing length function, resource utilization function and bottleneck function; An ant colony optimization module, which is used to construct a scheduling optimization objective function and a path heuristic function according to the simulation data set, and execute a search to generate an optimal material feeding scheduling sequence; A scheduling execution module, which is used to execute the scheduling process according to the optimal material feeding scheduling sequence, and control the equipment feeding according to the process sequence, resource allocation and time constraint; A feedback control module, which is used to monitor the execution deviation, and transmit the latest operation status back to the discrete event modeling module to trigger scheduling reconstruction, so as to realize the dynamic closed-loop optimization management of the concrete mixing plant production.

[0022] A production management method for a concrete mixing plant, which is applied to a production management system of a concrete mixing plant, and includes the following steps: S1. Construct a discrete event simulation model of the production process of the concrete mixing plant. The discrete event simulation model models the discrete events of the key processes of concrete mixing, loading, transportation and unloading, and defines the triggering conditions, execution order and resource constraints of each discrete event; S2. Collect the real-time operation data set of the concrete mixing plant. The real-time operation data set includes the operation status of each device, the arrival time of transport vehicles, the loading completion time and the number of queuing vehicles, and use the collected real-time operation data set as the input data of the discrete event simulation model; S3. Input the real-time operation data set into the discrete event simulation model, execute the discrete event simulation, and generate simulation data reflecting the material feeding process of the concrete mixing plant. The simulation data covers queuing status, equipment utilization rate and potential scheduling bottleneck information; S4. Initialize the ant colony optimization algorithm based on the simulation data, and at the same time define the scheduling optimization objective. Use the initialized ant colony optimization algorithm to perform scheduling solution based on the simulation data, and generate an optimal material feeding scheduling sequence through iterative search; S5. Apply the optimal material feeding scheduling sequence to the scheduling execution module of the concrete mixing plant to control the execution of the actual feeding process, so that the actual feeding scheduling is strictly carried out in the optimized order, and realize the management of the material feeding queue of the concrete mixing plant; S6. When it is detected that there is a significant deviation between the actual operation status and the simulation data, automatically feedback the newly collected real-time operation data set to the discrete event simulation model, update the operation status of the discrete event simulation model, and restart the closed-loop scheduling optimization process from step S4 to step S5 to realize the dynamic real-time adjustment of the material feeding queue control of the concrete mixing plant.

[0023] In this embodiment, S1 includes the following steps: S11. Set the set of key processes of the concrete mixing plant. The set of key processes includes the mixing process, the loading process, the transportation process, and the unloading process. The set of key processes serves as the basic process unit in the discrete event simulation model; S12. Define the discrete event simulation model as a structured modeling unit containing five components. The five components include the set of key processes, the set of resources, the set of process execution dependencies, the process event data set, and the set of process trigger functions. The five components together constitute the discrete event simulation model for simulating the process behavior in the material distribution process of the concrete mixing plant; Define the discrete event simulation model as a structured modeling unit for simulating the material distribution process of the concrete mixing plant. The model is represented as a five-tuple , where: represents the set of key processes, which defines all process nodes including mixing, loading, transportation, and unloading; represents the set of resources, including resources such as equipment and vehicles required for the corresponding processes and their capacity limits; represents the set of process execution dependencies, which is defined as the edges between nodes in a directed graph and is used to represent the sequential relationship between processes; represents the process event data set, which contains the time attributes (arrival, start, completion) of each process and the resource occupancy identifier; represents the set of process trigger functions, and each function corresponds to the process , and is used to determine whether the current simulation clock meets the start condition and whether the resources meet the allocation condition. If both are met, the process event is activated.

[0024] The discrete event simulation model runs in an event-driven mechanism during the simulation process, manages the event queue according to time priority, and determines whether a process is executable by sequentially calling the trigger function , combines the set of resources to allocate actual execution resources, and drives the system state to advance forward.

[0025] S13. Construct a process execution dependency graph. In the process execution dependency graph, the execution of each process depends on the completion of the previous process. If the execution of process A depends on the completion of process B, then record the dependency relationship from process B to process A in the set of process execution dependencies. The process execution dependency graph is used to constrain the timing sequence of process triggering; S14. Set up a resource set for the concrete mixing plant. The resource set includes multiple types of resources such as mixing equipment, loading systems, transport vehicles, and unloading platforms. Each type of resource has a unique resource identifier and a maximum available capacity, and the maximum available capacity is used to determine whether the resources required during the execution of a process meet the allocation conditions; S15. Construct a process event data set. Each data record in the process event data set includes process type, process arrival time, process start time, process end time, and the resource identifier occupied. The process event data set is used to drive the change process of the process state in the discrete event simulation model; S16. Define a process trigger function set for each process. The process trigger function set is used to determine whether the current system simulation time reaches the moment when the process is allowed to be triggered, and to determine whether the required resources have available capacity. If both the time condition and the resource condition are met, the output of the process trigger function is triggerable, otherwise it is non-triggerable; S17. At any system simulation time, construct the current triggerable process queue according to the results of the process trigger function. The triggerable process queue contains all processes that meet the trigger conditions at the current time. Define a process priority function, which is calculated based on the cumulative waiting time of the process, the idle time of the required resources, and the urgency of the process. Sort the processes in the triggerable process queue according to the results of the process priority function, and preferentially schedule the processes ranked higher to achieve coordinated scheduling of resource conflicts; The process priority function is a function used to sort multiple schedulable processes that meet the trigger conditions at the same simulation time, aiming to determine which processes should be preferentially selected to enter the execution state when resources are limited or there are conflicts. Essentially, the process priority function is a multi-index comprehensive scoring function, which assigns a priority score to the current process according to its operating environment and urgency. The higher the value, the higher the scheduling priority.

[0026] In the present invention, the process priority function has the general expression: ; where, is the priority score of process , and the larger the value, the earlier the process should be scheduled for execution. is the cumulative waiting time of process , that is, the current system simulation time minus the time when it enters the queue. is the idle time of the required resource type of process since the last release, that is, the current resource idle duration. is the urgency score of process , which is preset according to business rules or on-site weights (such as whether it is a particularly urgent order). is the weight coefficient of three parameters, which is set through historical data training or expert experience.

[0027] Cumulative waiting time From the process event data set The arrival time of each process and the current simulation time , and the calculation method is ; Resource idle time From the resource set , the timestamp when the resource was last released and the current simulation time The difference between ; Process urgency Set as: Standard process: ; Priority process: ; Urgent process: ; The urgency level can be set according to the actual task type or external instructions.

[0028] S18. Set a concurrent execution time window for each process. The concurrent execution time window defines the earliest start time and the latest completion time allowed for the process in the system simulation. In each system simulation cycle, if the time windows of two or more processes overlap, the required resources have no conflict, and the sequence constraint in the process execution dependency graph is not violated, then such processes are classified into the concurrent process set, and the processes in the concurrent process set are allowed to advance and execute simultaneously.

[0029] In this embodiment, a discrete event simulation model of the concrete mixing plant production process is constructed. The discrete event simulation model models the discrete events of the key processes of concrete mixing, loading, transportation, and unloading, and defines the trigger conditions, execution order, and resource constraints of each discrete event. This modeling process can structurally express the key processes in the actual production process in the form of events, enabling the system to have time-sequence logic controllability and resource allocation constraint, providing a unified modeling framework and dynamic simulation basis for subsequent queuing behavior simulation and scheduling optimization, and thus effectively supporting the global behavior prediction and bottleneck location analysis of the complex material distribution process of the concrete mixing plant.

[0030] In this embodiment, S2 includes the following steps: S21. Set the real-time operation data collection items of the concrete mixing plant. The real-time operation data collection items include three categories: process status data, resource status data, and scheduling queuing data. The three categories of data are respectively used to describe the process operation progress, resource availability, and scheduling execution pressure in the system; S22. The process status data includes the arrival time, start time, completion time, and execution status identifier of each process in the actual system. The process execution status identifier is used to indicate whether the process is in a waiting, executing, or completed state. This type of data is used to initialize or update the process event data set; S23. The resource status data includes the resource type, resource unique identifier, current occupancy status, occupancy start time, estimated release time, and maximum capacity. The resource status data is used to update the resource set in the discrete event simulation model and the resource constraint conditions in the process trigger function; S24. The scheduling queue data includes the current queue length, waiting duration of each process in the queue, and estimated start time. The scheduling queue data is used to assist in judging the process queuing pressure, process priority, and bottleneck position, and is an important input basis in the simulation process; S25. Structurally combine various types of data in the collection items to form a real-time operation data set 。

[0031] In this embodiment, S3 includes the following steps: S31. Input the real-time operation data set into the discrete event simulation model, perform corresponding structure mapping between the real-time operation data set and the process event data set and update the process status according to the current simulation time of the system, which is used to drive the process trigger judgment; It is necessary to perform one-to-one structure mapping between the fields in the process event data set and the structure elements in the discrete event simulation model, specifically including: Process mapping relationship: Match the "process type" field in the real-time operation data set with the key process set P in the discrete event simulation model. Different types of process records are respectively mapped to P1, P2, P3, and P4 in the key process set to determine the process category to which the data record belongs; Event status mapping relationship: Map the "arrival time", "start time", and "completion time" of each process in the real-time operation data set to the event trigger time field in the process event data set as the input of the event time attribute of the process, and fill or update the execution status of the corresponding process at the current simulation time; Resource status mapping relationship: Map the "occupied resource number" field in the real-time operation data set to the resource unique identifier in the resource set R, and update the occupancy status of the resource at the current simulation time to ensure the correct execution of the resource constraint judgment logic in the process trigger function; Queuing status mapping relationship: Map the "current queuing position" field in the real-time operation data set to the queue index position in the triggerable process queue, which is used to update the system queuing status and participate in the calculation of the process priority function; S32. At the current simulation time of the system, perform an event-driven time-advancing simulation operation. Relying on the key process set, resource set, process execution dependency set, process event data set, and process trigger function set defined in the discrete event simulation model, collect all processes in the waiting state that do not meet the trigger conditions to form a queued process set; S33. Based on the queued process set, for each type of process conduct statistics to calculate the queuing quantity of this process type at the current simulation time, obtaining the process queuing length function . The process queuing length function is classified by process type and calculates the number of processes of this type that have not met the trigger conditions at the current system simulation time, which is used to represent the waiting load status of different process types; Calculate the queuing quantity of this process type at the current simulation time. At the current system simulation time t, the simulation execution module sequentially determines the status of all processes to be executed according to the data records of each process in the process event data set. If a certain process meets the following two conditions: 1. The arrival time of the process is earlier than or equal to the current simulation time; 2. The process has not been triggered, that is, its status is "waiting", and the current resources are not allocable or the process execution dependency has not been met, then it is considered that the process is in the "queuing state". For a specific type of process, traverse all process records and filter out the total number of processes that are currently in the "queuing state" and the process type is equal to . This total number is the queuing quantity of this process type at the simulation time t.

[0032] Calculate the number of processes of this type that have not met the trigger conditions at the current system simulation time. At the current system simulation time t, based on the process event data set and the process trigger function, perform trigger judgment on all process instances of the process type , and count the number of process instances that have not simultaneously met the time trigger condition and the resource availability condition as the queuing length function value of this process type at time t .

[0033] S34. At the current simulation time, analyze the status of each type of resource in the resource set , calculate the quantity of this type of resource occupied by the executed processes, and combine it with the maximum available capacity of this type of resource to obtain the resource utilization rate function . The resource utilization rate function is composed of the ratio between the current occupied quantity of the resource and the maximum available capacity of the resource, which is used to measure the current operating load level of this type of resource; S35. Construct the process bottleneck function , the process bottleneck function takes the process queue length function at the current simulation time and the resource utilization rate function of the resource type corresponding to the process as inputs, and is constructed according to the ratio relationship between the two, and is used to reflect the degree of queue accumulation of the process when the resources are in a low utilization state. The larger the value of the process bottleneck function, the more serious the current restriction of the process on the system scheduling; S36. Within the total time period of the system simulation, perform cumulative average calculation on the time series of the process bottleneck function to form the process bottleneck index ; S37. Output the simulation data set , the simulation data set includes the system simulation time series, the process queue length function of each type of process, the resource utilization rate function of each type of resource, the process bottleneck function of each type of process, and the process bottleneck index. The simulation data set is used to reflect the operating states of each key process and resource in the concrete mixing plant's material delivery process driven by the current real-time operation data set.

[0034] In this embodiment, the process bottleneck index is based on the numerical sequence of the process bottleneck function within the entire simulation cycle. By averaging the bottleneck function values at all time nodes, a long-term evaluation index is obtained to reflect the degree of the process's constraint on the system operation in the entire material delivery process.

[0035] The present invention introduces a five-tuple structure including "key process set, resource set, process execution dependency set, process event data set, process trigger function set" into the modeling of the concrete mixing plant, realizes the granularity modeling of key processes such as mixing, loading, transportation, and unloading, can accurately determine the triggerable process queue according to each system simulation time, and combines resource availability and process priority function for dynamic sorting, realizes the full-process modeling and real-time calculation of process conflicts, concurrent scheduling, and bottleneck judgment. By constructing a simulation data set, simulation indexes such as process queue length function, resource utilization rate function, and bottleneck function are output, providing reliable global dynamic data support for subsequent optimization.

[0036] In this embodiment, S4 includes the following steps: S41. Initialize the parameter set of the ant colony optimization algorithm; S42. Construct a scheduling optimization objective function. The scheduling optimization objective function takes the process queue length function of various processes, the resource utilization rate function of each type of resource, and the process bottleneck function of each type of process in the simulation data set as inputs. The structure of the scheduling optimization objective function is used to comprehensively minimize the queuing waiting time, improve the resource load balance, and reduce the process bottleneck degree, and constitutes a global evaluation index for the scheduling path; The scheduling optimization objective function is a comprehensive function used to evaluate the quality of a material issuing scheduling sequence. It is the objective function in the entire ant colony optimization algorithm. In the present invention, the function is used to measure the comprehensive operation efficiency and scheduling balance of the system under a given scheduling scheme.

[0037] Scheduling optimization objective function With minimization as the optimization direction, it is mainly composed of three weighted parts: the waiting time of processes in the queue (represented by the process queue length function); the resource load balance (reflected by the resource utilization rate function); the severity of process bottlenecks (represented by the process bottleneck function).

[0038] It is defined as follows: ; Among them, is the scheduling path, is the number of process types, is the number of resource types, is the process 's average queue length, is the resource 's average utilization rate, is the average utilization rate of all resources, is the process 's bottleneck index, is the weight coefficient.

[0039] Process queue length function represents, during the simulation process, the number of processes of a certain type in the waiting queue that have not been triggered at time . The higher the value, the greater the waiting pressure of the process. Its construction method is generated by recording during the simulation process driven by the discrete event simulation model and is defined as: ; Among them, is the result of the process trigger function (0 means not triggered), that is, the current scheduling condition is not met.

[0040] Resource utilization rate function represents, at time , the current occupied ratio of resource type . It reflects whether the resource is overloaded or idle and is a key indicator for measuring the balanced scheduling of resources. The construction method is: ; Among them, is the number of resources currently being occupied, is the resource type 's maximum available capacity.

[0041] Process bottleneck function Used to measure the "bottleneck degree" of a certain type of process at time . It comprehensively considers the queuing accumulation degree of the process and the busy degree of the resources it depends on. The larger the value, the more likely the process is to become the bottleneck point of the system operation. Construction method: ; Among them, is the queuing length of the process at time , is the historical maximum queuing length of the process (normalized denominator), is the current utilization rate of the resources required by the process, is a small constant to prevent division by zero.

[0042] Summary of the relationship between the process queuing length function, the resource utilization rate function, and the process bottleneck function: Both the process queuing length function and the resource utilization rate function are derived from the time series simulation of the discrete event simulation process; The process bottleneck function is the combined result of the two and belongs to an advanced index; The scheduling optimization objective function takes these three as inputs and is a comprehensive scoring function for the global quality of the entire scheduling path, used to guide the ant colony to search for the optimal path.

[0043] S43. Construct a scheduling state space based on the queuing process set output in the simulation dataset. Each state in the scheduling state space corresponds to a set of legal process scheduling order arrangements, and the scheduling order arrangements must satisfy the process execution dependency constraint and the resource availability constraint; S44. Construct a path heuristic function. The path heuristic function consists of three parts, namely the process bottleneck index item, the process waiting time item, and the resource response time item. The bottleneck index item is set according to the bottleneck index value corresponding to the current process, the process waiting time item is set according to the cumulative waiting duration of the current process in the process queuing length function, and the resource response time item is set according to the idle response time calculated by the resource utilization rate function for the resource type required by the process. The path heuristic function is used to guide the ants to preferentially select the processes with less pressure on the system during the construction of the scheduling path; In the present invention, the path heuristic function is the key guiding factor for guiding the ants in the ant colony algorithm to select the next scheduling process when constructing the material issuing scheduling path. Its essence is a numerical function for evaluating whether each process to be scheduled is suitable to be preferentially executed in the current system state. The smaller the value of the path heuristic function, the lower the scheduling pressure of the current state of the process on the system and the more sufficient the resource preparation, and it is more suitable to be preferentially executed.

[0044] During the process of constructing the scheduling path, each ant will score and select all unscheduled operations based on the combined score of the path heuristic function and the pheromone concentration on the current path, so as to gradually and dynamically construct and optimize the entire scheduling path.

[0045] Path heuristic function For a certain candidate operation It is constructed and weighted by the following three parts: ; Among them, Is the operation bottleneck index item, indicating the degree of system bottleneck corresponding to this operation, which comes from the operation bottleneck index in the simulation dataset , the larger its value, the more likely this operation is to become a system operation bottleneck, Is the operation waiting time item, indicating the cumulative waiting time of this operation since it entered the queue, which can be obtained by accumulating the time of the operation queue length function in the simulation data, reflecting its scheduling urgency, Is the resource response time item, indicating the current response ability of the resource type required by this operation. Its value is the inverse function of the resource utilization rate function of the required resource type in the simulation data , the more idle the resource, the shorter the response time, and the smaller the value; , , Are the preset weighting factors in the scheduling strategy.

[0046] During the scheduling process, for each candidate operation that has not been included in the scheduling path , the system will calculate its path heuristic value based on the following three aspects: Degree of bottleneck: Referring to the bottleneck index value statistically obtained in the simulation, if this operation shows a serious bottleneck in the historical simulation, its corresponding heuristic value should be increased to avoid further increasing the system pressure; Queuing time: If this operation has waited for a long time, in order to reduce unfair scheduling and waiting backlog, its heuristic value should be appropriately reduced so that it is more likely to be scheduled first; Resource idle situation: If the resources required by the operation are currently relatively idle, then it can quickly obtain the execution conditions, and its heuristic value should be appropriately reduced to encourage it to be scheduled first.

[0047] Finally, the three indicators are combined according to the importance weights of the scheduling strategy to form a comprehensive score value , which is used to guide the ant to preferentially select the operation with the lowest score for expansion when constructing the path, so as to achieve double guarantees of path quality and scheduling stability.

[0048] S45. Initialize the ant colony search structure. Each ant randomly selects an operation from the scheduling state space as the starting node and gradually constructs a complete operation scheduling path according to the combined weight of the value calculated by the path heuristic function and the pheromone concentration. All ants generate a complete candidate solution respectively in each round of iteration. S46. After each round of iteration, score the paths of all ants according to the scheduling optimization objective function constructed in step S42, and update the pheromone concentration on the corresponding paths according to the scoring results. The update method includes two processes: pheromone intensification and pheromone evaporation. Pheromone intensification is enhanced positively according to the value of the scheduling optimization objective function, and pheromone evaporation is used to limit local over-concentration and avoid falling into local optimum. In each round of ant colony iteration, each ant independently constructs a scheduling path according to the path heuristic function and the pheromone matrix. The path represents the scheduling order of the operations. For example: This path corresponds to a specific operation scheduling order, which will be used to drive a complete discrete event simulation process later to obtain the system simulation running state corresponding to the path.

[0049] Basis for scoring: The scheduling optimization objective function defined in step S42 means that the scoring criterion is: the faster the path is executed, the more balanced the resources are, and the less severe the bottleneck is, the lower the score (the better).

[0050] Execution of scoring: Simulation - extraction - calculation process, for each ant The scoring process is as follows: Step 1: Apply the path to the discrete event simulation model for a complete simulation execution. Simulate operation triggering and queuing, and record the queuing length function at each time point , resource utilization rate function , bottleneck function ; Step 2: Average and normalize the above time series functions to obtain the average queuing length , average resource utilization rate , operation bottleneck index ; Step 3: Substitute into the scheduling optimization objective function to calculate the objective value ; Step 4: Record the scores of all ant paths for pheromone update (weaken pheromone for high - score paths and enhance pheromone for low - score paths); judge whether the optimal solution is reached; select the optimal path of this round.

[0051] Suppose there are two ants A and B: The path of A enables the system to run fast, utilize resources evenly, and have little queuing pressure. Then: ; If the path scheduling of B is improper, with large resource conflicts and obvious bottlenecks, then: ; Then the path of A is more optimal and will receive a pheromone reward with a higher weight.

[0052] Scoring the ant paths is equal to inputting the scheduling paths generated by each ant into the simulation model → extracting the performance indicators corresponding to the paths → substituting them into the objective function to calculate the path score value → for pheromone update and global optimal judgment.

[0053] S47. When the ant colony optimization algorithm reaches the set maximum number of iterations or the optimal path remains unchanged for several consecutive rounds, output the current optimal scheduling path as the optimal material feeding scheduling sequence generated under the current simulation state. The optimal material feeding scheduling sequence is used to guide the execution of the feeding process in the concrete mixing plant.

[0054] In this embodiment, the parameter set of the ant colony optimization algorithm includes the number of ants, the initial value of pheromone, the pheromone evaporation coefficient, the path selection control factor, and the maximum number of iterations. The number of ants is set according to the total number of queuing processes at the current simulation time, and the initial value of pheromone is set according to the process bottleneck index corresponding to each type of process Set, and the pheromone evaporation coefficient is set according to the average value of the overall resource utilization rate of the current system.

[0055] The present invention designs a composite path heuristic function composed of process bottleneck index items, process waiting time items, and resource response time items, and dynamically adjusts the pheromone evaporation coefficient and reinforcement mechanism according to the process queuing function and resource utilization function, enhancing the algorithm's perception ability of scheduling bottleneck nodes and effectively avoiding falling into a locally optimal scheduling sequence.

[0056] In this embodiment, S5 includes the following steps: S51. Input the optimal material feeding scheduling sequence into the scheduling execution module of the concrete mixing plant. The optimal material feeding scheduling sequence is composed of multiple scheduling units, and each scheduling unit includes a scheduling process identifier, a scheduling start time, the required resource number, and a resource allocation strategy; S52. According to the arrangement order of each scheduling unit in the optimal material feeding scheduling sequence, construct a feeding execution queue and load it into the scheduling queue control unit. The scheduling execution module activates the corresponding scheduling instruction when the following three scheduling conditions are met: All previous processes of the current scheduling process have been completed, meeting the process execution dependency set defined in the discrete event simulation model; The available capacity of the resources required for the scheduling process at the current moment in the resource set meets the resource allocation strategy; The system simulation time reaches or exceeds the scheduling start time corresponding to the scheduled process; S53. When the three scheduling conditions are simultaneously satisfied, the scheduling execution module sends a start instruction to the device corresponding to the specified resource number in the scheduling unit to control the execution of the actual process; if any condition is not satisfied, the scheduling unit pauses execution and enters the waiting state; S54. During the operation of the material issuing process, the scheduling execution module continuously detects whether the system execution status is consistent with the scheduling order. When it is found that there is a deviation between the execution order and the scheduling order, the sending of subsequent scheduling instructions is suspended, and the system scheduling status is rolled back to the position of the most recently completed scheduling process, waiting for the system status to realign with the scheduling logic; S55. Control the process scheduling management of the scheduling execution module of the concrete mixing plant according to the optimized scheduling order.

[0057] In this embodiment, by directly applying the optimal material issuing scheduling sequence to the scheduling execution module of the concrete mixing plant, a strict mapping relationship between the scheduling result and the on-site execution behavior is achieved, ensuring that each key process is executed in sequence according to the optimized order under the conditions of resource constraints and time dependencies, effectively avoiding problems such as process preemption, resource conflicts, and queuing jams. At the same time, a synchronous control and execution deviation rollback mechanism is introduced during the execution process, enabling the scheduling scheme to have fault tolerance for sudden changes, significantly improving the stability and continuity of the material issuing process, reducing the system waiting time and equipment idling rate, enhancing the controllability and practical executability of the scheduling process, and laying a stable execution foundation for subsequent closed-loop optimization and dynamic adjustment.

[0058] In this embodiment, S6 includes the following steps: S61. During the actual execution of the material issuing process of the concrete mixing plant, collect the current actual operation status data set for state comparison and analysis with the process queuing length function, resource utilization function, and process bottleneck function included in the simulation data set ; S62. Define the actual operation status data set . Compare the actual operation status data set with the simulation data set process by process and resource by resource. If any of the following conditions is met, it is determined that the system status deviation is significant: The actual completion time of a certain process in the actual operation status data set is delayed by more than the set threshold of the predicted completion time in the simulation data; The continuous occupation duration of a certain resource in the actual operation status data set is greater than the upper limit of the corresponding resource occupation time in the simulation data; The queuing time continuously exceeds the sum of the historical mean and standard deviation of the queuing length function in the simulation state in multiple consecutive processes; S63. When it is determined that the system deviation is greater than the threshold, immediately trigger the scheduling feedback mechanism and write the current system state into the latest real-time operation data set. ; S64. Input the latest real-time operation data set into the discrete event simulation model, reconstruct the simulation input state, execute the discrete event simulation process, and generate a new simulation data set. , and update the process queue length function, resource utilization function, and process bottleneck function in the simulation data. S65. Based on the updated simulation data set , re-initialize the ant colony optimization, perform path search and output the optimal scheduling path process to obtain a new optimal material feeding scheduling sequence. S66. Use the new optimal material feeding scheduling sequence as a feedback scheduling instruction and re-enter it into the concrete mixing plant scheduling execution module to re-manage the process execution, replace the old scheduling path, and achieve dynamic closed-loop optimization management of the concrete mixing plant production.

[0059] The present invention establishes a complete set of scheduling deviation detection and feedback closed-loop optimization system. When it detects state deviations such as "process completion delay", "abnormal continuous resource occupation", or "multiple process continuous queuing exceeding the threshold", it can automatically trigger scheduling reconstruction. The system will remap the newly collected real-time operation data set to the simulation model, reconstruct the simulation state in real time, and start the ant colony algorithm for path re-search, outputting a new optimal scheduling path to replace the old execution path, thereby achieving the continuity and dynamic optimality of production scheduling.

[0060] Example 1: In November 2024, a commercial concrete mixing plant with an annual production capacity of over 500,000 cubic meters located in County A continuously had problems such as material feeding delays, frequent resource conflicts, and queuing vehicle congestion during daily operations. The mixing plant has 5 JS1000 type mixing main engines, 3 sets of automatic loading systems, 12 concrete transport vehicles, and 2 unloading platforms, with a daily material feeding batch exceeding 180 times. Limited by the original scheduling system which only supports static priority strategies at fixed time intervals, the equipment utilization rate is low and the resource allocation logic cannot be dynamically adjusted.

[0061] In the operation log on October 25, 2024, during the period from 13:00 to 15:00, 4 consecutive events of exceeding the standard of material delivery waiting time occurred. In the most serious case, the vehicle queue length reached 9 vehicles, and the average waiting time per vehicle exceeded 31 minutes, significantly affecting the construction rhythm at the project site. Dispatcher Engineer Zhang manually adjusted the priority twice on the dispatching system interface, but still could not effectively solve the dispatching lag problem caused by the conflict of mixing equipment resources and the sudden increase in queuing vehicle trips. Based on this background, this station decided to introduce the present invention and deploy it in the actual production environment for a 7-day comparative test to verify the practicability and improvement effect of the system.

[0062] The system was launched on November 3, 2024, and discrete event simulation modeling was carried out first. The technical team defined four types of key processes according to the actual production process of the mixing plant: mixing, loading, transportation, and unloading, which were respectively marked as MIX, LOAD, TRANS, and DISCHARGE. During the initial modeling stage, the operation log data from October 20 to November 2, 2024 was collected, including the start and end times of operations, equipment IDs, material delivery sequences, and vehicle dispatching schedules, constituting the basic process event data set, with the number of recorded entries reaching 4,286.

[0063] Subsequently, a resource set was set through the modeling interface, including mixing equipment numbered from SB001 to SB005, loading ports numbered from ZL001 to ZL003, transport vehicles numbered from VC001 to VC012, and two unloading platforms DP001 and DP002. The maximum available capacity, operation cycle, and cooling interval of each type of resource were defined, and a resource usage mapping table was established. The system automatically generated a process execution dependency graph to clarify the sequence and constraint logic between each process. In Example 1, loading must wait for mixing to be completed and the corresponding mixing equipment status to be available, and the transportation task must be bound to an idle vehicle and the destination site to be able to receive unloading.

[0064] Real-time operation data was collected in real time by the industrial computing gateway deployed in the central control room and synchronized with the dispatching host through the OPC protocol. The data collection frequency was 1Hz. The simulation engine updated the system status every 20 seconds, calculated the current queue length function qk(t), resource utilization function uj(t), and bottleneck function βk(t) indicators, and generated a snapshot Dsim of the simulation data set once an hour for the dispatching optimization module to call.

[0065] In the scheduling optimization module, the adaptive ant colony algorithm designed by the present invention starts to perform path search. Taking 9:00 on November 4th as an example, when the system simulation status shows that the queuing length function of the mixing process qMIX(t) = 7, the loading resource ZL002 is in an idle state, the utilization rate of the transport vehicle uTRANS = 87.5%, and the bottleneck function βMIX(t) = 0.63, at this time, the path heuristic function automatically determines to preferentially schedule two tasks, MIX-4 and MIX-7, in the MIX→LOAD→TRANS path, and allocates SB003 and ZL002 as the corresponding devices to control the adjustment of the order of the material issuing queue. This strategy effectively avoids the "non-bottleneck priority coverage problem" existing in the original scheduling system, that is, low-waiting processes are frequently preempted by high-waiting processes for resources.

[0066] During the advancement of the system simulation, the scheduling execution module automatically verifies three scheduling conditions (completion of the previous process, availability of resources, arrival at the scheduling time), and sends a start instruction to the PLC system through the OPC-UA interface. At 13:45 on November 4th, the mixing process MIX-12 corresponding to the scheduling instruction DI202411041345-A was successfully triggered, the resource allocation was SB001, the loading port was ZL003, the estimated completion time was 13:58, and the actual completion time was 13:57, with the error controlled within 60 seconds.

[0067] At the same time, to verify the feedback reconstruction ability of the system under emergencies, the technical team deliberately delayed the arrival time of transport vehicle VC004 by 8 minutes in the morning of November 6th, triggering the delay phenomenon of process DISCHARGE-22. The system successfully detected that the deviation between the actual process completion time and the simulation data reached the standard threshold (the delay exceeded 115 seconds predicted by the simulation), automatically triggered the simulation model update and path reconstruction mechanism. After replacing the old scheduling path, the cumulative delay time of subsequent tasks was controlled within 3 minutes, effectively avoiding resource accumulation and scheduling deadlocks.

[0068] The following is a summary of the key comparison data between the system of the present invention and the original fixed priority strategy within a 7-day cycle (from November 3rd to November 9th, 2024): Table 1 Comparison of project data between the original scheduling system and the system of the present invention Project Original scheduling system System of the present invention Improvement range Average operation waiting time 18.7 minutes 11.2 minutes ↓40.1% Utilization rate of mixing equipment 63.5% 81.9% ↑28.9% Frequency of loading resource conflicts 42 times per day 17 times per day ↓59.5% Number of delays in material issuing tasks 23 times per day 6 times per day ↓73.9% Scheduling response time 2.7 minutes 1.3 minutes ↓51.9% Variance of process bottleneck index 0.098 0.042 ↓57.1% In terms of training samples, the algorithm of the present invention adopts a simulation data-driven path reinforcement mechanism. The cumulative number of training path samples is 3,761, and each sample includes a scheduling order, a resource allocation combination, a system bottleneck state vector, and a final execution feedback result. Among them, the number of effective training samples is 3,212, accounting for 85.4%, and the F1 score of the sample quality evaluation reaches 0.936. Compared with the original system that only relies on a limited rule base (about more than 1,000 fixed policy combinations) under the historical priority strategy, the system of the present invention can significantly expand the policy space and realize self-learning and optimization from actual feedback.

[0069] Through real deployment and data comparison verification, Example 1 fully proves the feasibility and significant advantages of the present invention in actual production scenarios. In the scenarios of high-frequency scheduling of multiple devices and intensive conflicts during the peak material delivery period, the simulation modeling and adaptive ant colony scheduling mechanism constructed by the present invention has stronger scheduling flexibility and resource coordination capabilities, effectively solving the problems of lagging resource bottleneck identification, frequent task conflicts, and rigid path strategies existing in the original system, and having good engineering practicability and popularization value.

[0070] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A concrete mixing station production management system, characterized in that: Includes the following modules: Discrete event modeling module, used to build a discrete event simulation model, which includes a set of key processes, a set of resources, a set of process execution dependencies, a set of process event data, and a set of process trigger functions; A data acquisition module is used to obtain the real-time operation data set of the concrete mixing plant and use the real-time operation data set as input to the discrete event simulation model; A simulation execution module is used to execute an event-driven simulation process and generate a simulation data set including a process queue length function, a resource utilization function, and a bottleneck function; The ant colony optimization module is used to construct the scheduling optimization objective function and path heuristic function according to the simulation data set, and perform search to generate the optimal material delivery scheduling sequence; The scheduling execution module is used to execute the scheduling process according to the optimal material delivery scheduling sequence and control the equipment delivery according to the process sequence, resource allocation and time constraints; The feedback control module is used to monitor execution deviations and transmit the latest operating status back to the discrete event modeling module to trigger scheduling reconstruction and realize dynamic closed-loop optimization management of concrete mixing station production.

2. A concrete mixing station production management method, applied to a concrete mixing station production management system according to claim 1, characterized in that: The following steps are involved: S1. Construct a discrete event simulation model of the concrete mixing plant production process. The discrete event simulation model models the discrete events of the key processes of concrete mixing, loading, transportation and unloading, and defines the triggering conditions, execution sequence and resource constraints of each discrete event; S2. Collect the real-time operation data set of the concrete mixing station, which includes the operation status of each equipment, the arrival time of the transport vehicle, the loading completion time and the number of queued vehicles; S3. Input the real-time operation data set into the discrete event simulation model, perform discrete event simulation, and generate simulation data reflecting the material delivery process of the concrete mixing station. The simulation data includes queue status, equipment utilization rate, and potential scheduling bottleneck information; S4. Initialize the ant colony optimization algorithm based on the simulation data, define the scheduling optimization goal, use the initialized ant colony optimization algorithm to perform scheduling solution based on the simulation data, and generate the optimal material delivery scheduling sequence through iterative search; S5. Apply the optimal material delivery scheduling sequence to the scheduling execution module of the concrete mixing station to control the execution of the actual delivery process so that the actual delivery scheduling is strictly carried out in the order generated by the optimization; S6. When a significant deviation is detected between the actual operating status and the simulation data, the newly collected real-time operating data set is automatically fed back to the discrete event simulation model, the operating status of the discrete event simulation model is updated, and the closed-loop scheduling optimization process from step S4 to step S5 is restarted.

3. A concrete mixing station production management method according to claim 2, characterized in that: The S1 comprises the following steps: S11. Set a set of key processes of a concrete mixing plant, which includes a mixing process, a loading process, a transportation process and an unloading process. The key process set is used as a basic process unit in a discrete event simulation model. S12. A discrete event simulation model is defined as a structured modeling unit comprising five components, wherein the five components include a key process set, a resource set, a process execution dependency set, a process event data set, and a process trigger function set; S13. Construct a process execution dependency graph, in which the execution of each process depends on the completion of the previous process. If the execution of process A depends on the completion of process B, then the dependency from process B to process A is recorded in the process execution dependency set. The process execution dependency graph is used to constrain the timing sequence of process triggering; S14. Set a resource set for a concrete mixing station. The resource set includes multiple types of resources, including mixing equipment, loading systems, transport vehicles, and unloading platforms. Each type of resource has a unique resource identifier and a maximum available capacity. The maximum available capacity is used to determine whether the resources required for the execution of the process meet the allocation conditions. S15. Construct a process event data set. Each data record in the process event data set includes the process type, process arrival time, process start time, process end time, and the resource identifier occupied. The process event data set is used to drive the process state change process in the discrete event simulation model; S16. Define a process trigger function set for each process. The process trigger function set is used to determine whether the current system simulation time reaches the time when the process is allowed to be triggered, and whether the required resources have available capacity. If both the time condition and the resource condition are met, the process trigger function output is triggerable, otherwise it is not triggerable; S17. At any system simulation moment, a current triggerable process queue is constructed based on the process trigger function result. The triggerable process queue includes all processes that meet the trigger conditions at the current moment. A process priority function is defined. The process priority function is calculated based on the cumulative waiting time of the process, the idle time of the required resources, and the urgency of the process. The processes in the triggerable process queue are sorted according to the process priority function result, and the processes with the highest sorting are preferentially scheduled to realize the coordinated scheduling of resource conflicts. S18. Set a concurrent execution time window for each process. The concurrent execution time window defines the earliest start time and the latest completion time allowed for the process in the system simulation. In each system simulation cycle, if the time windows of two or more processes overlap and the required resources do not conflict and do not violate the sequence constraints in the process execution dependency graph, then such processes will be included in the concurrent process set, and the processes in the concurrent process set will be allowed to be executed simultaneously.

4. A concrete mixing station production management method according to claim 3, characterized in that: The S3 comprises the following steps: S31. Real-time operational data sets Input into the discrete event simulation model to combine the real-time operation data set with the process event data set Carry out corresponding structure mapping, update the process status according to the current simulation time of the system, and drive the process trigger judgment; S32. At the current simulation moment of the system, an event-driven time advancement simulation operation is performed, and all processes in a waiting state and not meeting the triggering conditions are collected based on the key process set, resource set, process execution dependency set, process event data set and process trigger function set defined in the discrete event simulation model to form a queued process set; S33. Based on the queued process set, each type of process Perform statistics and calculate the number of queues of this process type at the current simulation time to obtain the process queue length function ,The process queue length function takes the process type as the classification basis, and calculates the number of processes of this type that have not met the trigger conditions at the current system simulation time; S34. At the current simulation time, for each type of resource in the resource set The resource utilization function is obtained by analyzing the status of the resource and calculating the amount of resources occupied by the executed processes. ,The resource utilization function is composed of the ratio between the current ,occupancy of resources and the maximum available capacity of the resources; S35. Construct process bottleneck function The process bottleneck function takes the process queue length function at the current simulation time and the resource utilization function of the resource type corresponding to the process as input, and is constructed according to the ratio between the two. It is used to reflect the queue accumulation degree of the process when the resources are in a low utilization state. The larger the value of the process bottleneck function, the more serious the restriction of the process on the system scheduling. S36. Within the total system simulation time period, the time series of the process bottleneck function is cumulatively averaged to form a process bottleneck index ; S37. Output simulation data set ,The simulation data set includes the system simulation time series, the ,process queue length function of each type of process, the resource utilization ,function of each type of resource, the process bottleneck function of each type of ,process, and process bottleneck index. The simulation data set is used to reflect the ,operation status of each key process and resource in the material ,dispatching process of a concrete mixing station driven by the current ,real-time operation data set.

5. A concrete mixing station production management method according to claim 4, characterized in that: The process bottleneck index is based on the numerical sequence of the process bottleneck function in the entire simulation cycle. By averaging the bottleneck function values ​​at all time nodes, the long-term evaluation index is obtained to reflect the degree of constraint of the process on the system operation in the entire material delivery process.

6. A concrete mixing station production management method according to claim 4, characterized in that: The S4 comprises the following steps: S41. Initialize the ant colony optimization algorithm parameter set; S42. Construct a scheduling optimization objective function, which takes the process queue length function of each type of process in the simulation data set, the resource utilization function of each type of resource, and the process bottleneck function of each type of process as input. The scheduling optimization objective function structure is used to comprehensively minimize the queue waiting time, improve the resource load balance and reduce the process bottleneck degree, forming a global evaluation index for the scheduling path; S43. Construct a scheduling state space based on the queued process set output from the simulation data set. Each state in the scheduling state space corresponds to a set of legal process scheduling sequences. The scheduling sequence must satisfy the process execution dependency constraints and resource availability constraints. S44. Construct a path heuristic function. The path heuristic function consists of three parts, namely, a process bottleneck index item, a process waiting time item, and a resource response time item. The bottleneck index item is set according to the bottleneck index value corresponding to the current process, the process waiting time item is set according to the accumulated waiting time of the current process in the process queue length function, and the resource response time item is set according to the idle response time calculated in the resource utilization function based on the resource type required by the process; S45. Initialize the ant colony search structure. Each ant randomly selects a process from the scheduling state space as the starting node, and gradually builds a complete process scheduling path according to the joint weight of the path heuristic function calculation value and the pheromone concentration. All ants generate a complete candidate solution in each round of iteration. S46. After each round of iteration, the paths of all ants are scored according to the scheduling optimization objective function constructed in step S42, and the pheromone concentration on the corresponding path is updated according to the scoring result. The updating method includes two processes: pheromone enhancement and pheromone volatilization. The pheromone enhancement is positively feedback enhanced according to the scheduling optimization objective function value, and the pheromone volatilization is used to limit local over-concentration and avoid falling into the local optimum. S47. When the ant colony optimization algorithm reaches the set maximum number of iterations or the optimal path remains unchanged in several consecutive rounds, the current optimal scheduling path is output as the optimal material delivery scheduling sequence generated under the current simulation state. The optimal material delivery scheduling sequence is used to guide the execution of the delivery process of the concrete mixing station.

7. A concrete mixing station production management method according to claim 6, characterized in that: The ant colony optimization algorithm parameter set includes the number of ants, the initial value of pheromone, the volatility coefficient of pheromone, the path selection control factor and the maximum number of iterations. The number of ants is set according to the total number of queued processes at the current simulation time, and the initial value of pheromone is set according to the process bottleneck index corresponding to each type of process. The pheromone volatilization coefficient is set according to the average value of the overall resource utilization of the current system.

8. A concrete mixing station production management method according to claim 7, characterized in that: The S5 comprises the following steps: S51, inputting the optimal material delivery scheduling sequence into the concrete mixing station scheduling execution module, the optimal material delivery scheduling sequence is composed of multiple scheduling units, each scheduling unit includes a scheduling process identifier, a scheduling start time, a required resource number and a resource allocation strategy; S52. According to the arrangement order of each scheduling unit in the optimal material dispatch scheduling sequence, a dispatch execution queue is constructed and loaded into the dispatch queue control unit. The dispatch execution module activates the corresponding dispatch instruction when the following three dispatch conditions are met: All the predecessor processes of the currently scheduled process have been completed, satisfying the process execution dependency set defined in the discrete event simulation model; The available capacity of the resources required for the scheduling process in the resource set at the current moment satisfies the resource allocation strategy; The system simulation time reaches or exceeds the scheduling start time corresponding to the scheduling process; S53, when the three scheduling conditions are met at the same time, the scheduling execution module sends a start instruction to the equipment corresponding to the resource number specified in the scheduling unit to control the actual process execution; if any condition is not met, the scheduling unit suspends execution and enters a waiting state; S54, the scheduling execution module continuously detects whether the system execution state is consistent with the scheduling sequence during the running of the material delivery process. When it is found that there is a deviation between the execution sequence and the scheduling sequence, the sending of subsequent scheduling instructions is suspended, and the system scheduling state is rolled back to the position of the most recently completed scheduling process, waiting for the system state to be re-aligned with the scheduling logic; S55, controlling the optimized scheduling sequence of the concrete mixing station scheduling execution module to perform process scheduling management.

9. A concrete mixing station production management method according to claim 7, characterized in that: The S6 comprises the following steps: S61. During the actual execution of the concrete mixing station delivery process, the current actual operation status data set is collected for comparison with the simulation data set. The process queue length function, resource utilization function and process bottleneck function contained in the process queue length function, resource utilization function and process bottleneck function are compared and analyzed; S62. Define the actual operation status data set , compare the current actual operation status data set with the simulation data set process by process and resource by resource. If any of the following conditions is met, the system status deviation is considered significant: A data set of the actual operation status of a process The actual completion time in the simulation data is delayed beyond the set threshold of the predicted completion time; A resource in the actual running state data set The continuous occupation time in is greater than the corresponding resource occupation time upper limit in the simulation data; The queue time is continuously greater than the sum of the historical mean and standard deviation of the queue length function under the simulation state in multiple consecutive processes; S63: When the system deviation is greater than the threshold, the dispatch feedback mechanism is triggered immediately to write the current system status into the latest real-time operation data set. ; S64, the latest real-time operational data set Input to the discrete event simulation model, reconstruct the simulation input state, execute the discrete event simulation process, and generate a new simulation data set , update the process queue length function, resource utilization function and process bottleneck function in the simulation data; S65, based on the updated simulation data set , re-initialize the ant colony optimization, path search and optimal scheduling path output process to obtain a new optimal material delivery scheduling sequence; S66. Re-input the new optimal material delivery scheduling sequence as the feedback scheduling instruction into the scheduling execution module of the concrete mixing station, re-perform the process execution management, and replace the old scheduling path.

Images