CN114611401A - Multi-level complex service intelligent simulation method and system - Google Patents
Multi-level complex service intelligent simulation method and system Download PDFInfo
- Publication number
- CN114611401A CN114611401A CN202210269296.1A CN202210269296A CN114611401A CN 114611401 A CN114611401 A CN 114611401A CN 202210269296 A CN202210269296 A CN 202210269296A CN 114611401 A CN114611401 A CN 114611401A
- Authority
- CN
- China
- Prior art keywords
- node
- level
- service
- task
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000004088 simulation Methods 0.000 title claims abstract description 76
- 238000000034 method Methods 0.000 title claims abstract description 58
- 238000013528 artificial neural network Methods 0.000 claims abstract description 86
- 238000012545 processing Methods 0.000 claims abstract description 73
- 230000003993 interaction Effects 0.000 claims abstract description 55
- 230000008569 process Effects 0.000 claims abstract description 36
- 230000000694 effects Effects 0.000 claims abstract description 18
- 239000013598 vector Substances 0.000 claims description 69
- 230000015654 memory Effects 0.000 claims description 28
- 230000002452 interceptive effect Effects 0.000 claims description 17
- 238000010606 normalization Methods 0.000 claims description 12
- 238000007726 management method Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 238000004140 cleaning Methods 0.000 claims description 6
- 230000007787 long-term memory Effects 0.000 claims description 6
- 230000006403 short-term memory Effects 0.000 claims description 4
- 238000004891 communication Methods 0.000 claims description 3
- 239000003795 chemical substances by application Substances 0.000 description 8
- 230000006399 behavior Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/445—Program loading or initiating
- G06F9/44505—Configuring for program initiating, e.g. using registry, configuration files
- G06F9/4451—User profiles; Roaming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Entrepreneurship & Innovation (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Marketing (AREA)
- Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a multi-level complex service intelligent simulation method and a multi-level complex service intelligent simulation system. The method comprises the following steps: describing the hierarchical relationship of the complex service to determine the superior-inferior relationship of each node in each hierarchy; determining the node attribute of each node to obtain the task content composition of each node; constructing a node local level service processing model based on a deep neural network to obtain the total time consumed by the local level node to complete a superior task; constructing a service model of upper-level and lower-level interaction to obtain the total time consumption of the lower-level node for completing the upper-level task; setting configuration parameters of each node according to multi-level complex service conditions; calling a node current-level service processing model and a service model of upper and lower level interaction for simulation, and recording time data of each node completing service activity in the simulation process; adjusting the configuration parameters of each node, and performing simulation again; and after the simulation process is finished, comparing the simulation time data under different configuration parameters to obtain the optimal configuration parameters.
Description
Technical Field
The invention relates to the technical field of complex service modeling and the technical field of artificial intelligence, in particular to a multi-level complex service intelligent simulation method and system.
Background
Business Process Management (BPM), which mainly comprises the steps of: determining a business process existing in an organization; modeling a business process and compiling a process list; deploying the data to a process engine, and determining evaluation indexes such as efficiency indexes, benefit indexes and result indexes; executing the flow, and outputting a quantitative evaluation result; the boot flow improves the actions. The enterprise business process management provides a business process evaluation and improvement method from the overall perspective, generally aims at single-level, multi-horizontal department and multi-step business activities, and is not suitable for multi-level complex business activities.
Complex business activities such as task assignment, action monitoring, information feedback and the like are carried out between upper and lower levels in a plurality of levels under the scenes such as epidemic situation prevention and control, earthquake relief, flood fighting, emergency rescue, traffic transportation command, industrial control scheduling and the like. The business activity data is continuously generated along with the business activity, and the distribution, execution and feedback of the data have the characteristics of multiple levels, high timeliness, high frequency, high authority and the like. Traditionally, work distribution and staffing of these business activities are mainly evaluated by human subjectivity, and there is a lack of deduction and demonstration in advance and a lack of data support afterwards. The adoption of physical equipment, an actual network environment and manpower for large-scale service simulation can cause the consumption of a large amount of manpower and material resources, and the consumption is expensive.
Disclosure of Invention
In order to solve the problem that the prior Business Process Management (BPM) is multi-directed to a single-level business process and the collaborative effect of a plurality of complex business multi-level multi-entities is difficult to evaluate, the invention provides a multi-level complex business intelligent simulation method and a multi-level complex business intelligent simulation system.
In one aspect, the present invention provides a multi-level complex service intelligent simulation method, including:
step 1, taking each service entity in the complex service as a node, and describing the hierarchical relationship of the complex service to determine the superior-inferior relationship of each node in each hierarchy;
step 2, determining the node attribute of each node according to the superior-inferior relation of each node to obtain the task content composition of each node;
step 3, constructing a node level business processing model based on a deep neural network, and obtaining the total time consumption of the node level to finish a superior task by utilizing the node level business processing model;
step 4, constructing a service model of the upper-level and lower-level interaction, and obtaining the total time consumption of the lower-level node for completing the upper-level task by using the service model of the upper-level and lower-level interaction;
step 5, setting configuration parameters of each node according to multi-level complex service conditions;
step 6, calling the node current-level service processing model and the upper and lower-level interactive service model for simulation according to the given configuration parameters, and recording time data of each node completing service activities in the simulation process;
step 7, adjusting configuration parameters of each node, and executing step 6 again;
and 8, comparing simulation time data under different configuration parameters after the simulation process is finished to obtain the optimal configuration parameters.
Further, in step 2, the node attributes include a top level node, a middle level node and a bottom level node; in a corresponding manner, the first and second electrodes are,
the task content constitution of the top level node comprises the following steps: processing the current-level service and assigning lower-level tasks;
the task content constitution of the middle level node comprises the following steps: receiving a superior task, processing a current-level service and dispatching a subordinate task;
the task content of the bottommost node comprises the following steps: receiving superior tasks and processing the current-level service.
Further, in step 3, the neural network structure adopted by the node current-level service processing model includes a first long-term and short-term memory neural network, a second long-term and short-term memory neural network and a deep full-link neural network;
the training process of the node current-level business processing model specifically comprises the following steps:
cleaning the service data of the node to obtain an atom work vector, a worker vector and the normalization time for completing the current-level service; the atomic work is the minimum work unit forming the current-level service, the atomic work vector is obtained by performing representation description on the atomic work through the feature vector, and the staff vector is obtained by performing representation description on staff participating in the current-level service through the feature vector; one atomic job corresponds to one atomic job vector and one worker corresponds to one worker vector;
respectively taking an atomic working vector and a working personnel vector as the input of a first long-short time memory neural network and a second long-short time memory neural network, cascading the output of the last unit of the first long-short time memory neural network and the output of the last unit of the second long-short time memory neural network, inputting the cascading vector into a deep fully-connected neural network, taking the normalization time of the completion of the current-level service as the output of the deep fully-connected neural network, and then adopting supervised learning neural network parameters or simulating the learning neural network parameters through expert data, wherein the trained neural network is the node current-level service processing model.
Further, step 3 specifically includes:
according to the normalized time output by the node current-level service processing model, the time actually spent by current-level service processing of the current-level node j is calculated according to the formula (1)
Wherein, TmaxRepresents the maximum completion time, T, of the current level of serviceminThe minimum completion time of the current-level service is represented, and y is the normalized time;
checking whether the current node j has a lower node: if the node j exists, the time actually spent by the current-level service processing of each lower-level node is calculated, and the time of the current-level node j depending on the lower-level node is calculated according to the formula (2)
If not, then order
Wherein,
representing the time actually spent by the current-level service processing of the lower-level node i of the current-level node j;
calculating according to a formula (3) to obtain the total time consumption of the current-level node j for completing the superior task
Further, the service model of the upper and lower level interaction specifically includes:
determining a task allocation mode so as to calculate task interaction time under the specific task allocation mode; the task allocation mode comprises an online interactive allocation task mode and an offline interactive allocation task mode;
according to the task interaction time and the total time consumption of the subordinate node i to finish the superior task
Calculated according to the formula (4)
Wherein,
indicating a task interaction time required from the current node j to inform its lower node i of the task ready to receive until the lower node i receives the task.
Further, if online interaction is used to assign tasks,
and calculating the mean value according to the receiving sequence notified by the node at the current level and the number of the subordinate nodes under the node at the current level, and calculating the variance according to the working state of the subordinate nodes and the stability of the communication link.
Further, if the task is distributed by using the offline interaction, the calculation is performed by using the formula (5)
Wherein, Ti,goto、Ti,receiveAnd Ti,gobackAnd respectively obtaining the task time and the route time for the subordinate node i to return to the position of the subordinate node according to the notice after the subordinate node i arrives.
Further, in step 5, the configuration parameters include an atomic work vector, a staff vector, and upper and lower level interaction parameters of the node.
In another aspect, the present invention provides a multi-level complex service intelligent simulation system, including: the system comprises a service level description module, a planned configuration module, a node current-level service processing simulation module, a node and lower-level interactive simulation module and a simulation progress management module;
the service level description module is used for taking each service entity in the complex service as a node and describing the level relationship of the complex service to determine the superior-inferior relationship of each node in each level; the system comprises a plurality of nodes, a task processing module and a task processing module, wherein the nodes are used for determining the node attribute of each node according to the superior-inferior relation of each node to obtain the task content composition of each node;
the planning configuration module is used for planning configuration parameters of each node according to multi-level complex service conditions; and adjusting configuration parameters of each node;
the node local-level service processing simulation module is used for calling a node local-level service processing model to simulate the local-level service processing process according to the given configuration parameters and obtaining the total consumption time of the local-level node for completing the superior task;
the node and lower interactive simulation module is used for calling a service model of upper and lower interaction to simulate the upper and lower service interaction process according to given configuration parameters and obtaining the total time consumption of the lower node for completing the upper task;
the simulation progress management module is used for starting, pausing or stopping a simulation process.
Further, the system further comprises: a node current-level service processing neural network training module;
the node level business processing neural network training module is used for cleaning the business data of the node to obtain an atomic work vector, a worker vector and the normalization time for completion of the level business; respectively taking an atomic working vector and a working personnel vector as the input of a first long-short time memory neural network and a second long-short time memory neural network, cascading the output of the last unit of the first long-short time memory neural network and the output of the last unit of the second long-short time memory neural network, inputting the cascading vector into a deep fully-connected neural network, taking the normalization time of the completion of the current-level service as the output of the deep fully-connected neural network, and then adopting supervised learning neural network parameters or simulating the learning neural network parameters through expert data, wherein the trained neural network is the node current-level service processing model.
The invention has the beneficial effects that:
the invention can perform analog simulation on the execution process of the multi-level complex service. By analyzing the task quantity of the business among all levels and the execution time of the business among all levels, a scientific basis is provided for the personnel allocation and time consumption of all levels of the whole business activity on the premise that the total number of people is fixed.
Drawings
Fig. 1 is a schematic flow chart of a multi-level complex service intelligent simulation method according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a deep neural network structure adopted by a node present-level service processing model according to an embodiment of the present invention;
fig. 3 is a structural block diagram of a multi-level complex service intelligent simulation system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
On the one hand, the multi-level complex business is difficult to evaluate the multi-level complex business activity by the benefit of the change of the state of the bottom level entity because the multi-level complex business does not directly relate to the transition of the state of the bottom level entity. On the other hand, the multi-level complex service depends on the working magnitude and characteristics of each node at each level and the process of upper and lower level interaction, and the total time consumption and efficiency are determined by the number and characteristics of workers under the condition of limited workers. Under the condition, a multi-level complex business activity modeling method is provided, and a simulation system is constructed to analyze the working allocation and the personnel allocation of the multi-level business by using the time consumption of the multi-level complex business as a basic evaluation standard.
Example 1
As shown in fig. 1, an embodiment of the present invention provides a multi-level complex service intelligent simulation method, including the following steps:
s101, taking each service entity in the complex service as a node, and describing the hierarchical relationship of the complex service to determine the superior-inferior relationship of each node in each hierarchy;
in particular, a data structure (typically, such as a tree data structure) may be employed to describe the hierarchical relationship of complex services.
For example, Dict _ units _ backbone { ' entity a ': entity B1', ' entity B2',
'entity B1' [ 'entity C1', 'entity C2', 'entity C3' ],
'entity B2' [ 'entity C4', 'entity C5' ] }
The tree data structure represents: entity a is the top node of the service, i.e. the top director, and its immediate subordinate nodes are entity B1 and entity B2, while B1 has three subordinate nodes, namely, subordinate entity C1, entity C2 and entity C3, and B2 has two subordinate nodes, namely, subordinate entity C4 and entity C5. It represents an entity tree of complex traffic, connected by directed edges from the root node of the tree to the leaf nodes. The number of edges in a path from the root to a certain node is the degree of the node, and the degree of the root is defined as 0. Nodes with the same degree are defined as corresponding level cooperative nodes; the degree difference is 1, and two nodes are directly connected by edges and are directly superior and inferior; the degree difference is larger than 1, but two nodes which are connected by a path from the root node exist, and the two nodes are an indirect upper level and a indirect lower level. The entity tree may be present in a json file.
S102, determining the node attribute of each node according to the superior-subordinate relation of each node to obtain the task content composition of each node;
specifically, the node attributes include a top level node, a middle level node, and a bottom level node; in a corresponding manner, the first and second electrodes are,
the task content constitution of the top level node comprises the following steps: processing the current-level service and assigning lower-level tasks; the task content constitution of the middle level node comprises the following steps: receiving a superior task, processing a current-level service and dispatching a subordinate task; the task content of the bottommost node comprises the following steps: receiving the superior task and processing the current service.
For example, entity a is the top level node, entity B1 and entity B2 are the middle level nodes, and entity C1, entity C2, entity C3, entity C4, and entity C5 are the bottom level nodes.
It should be noted that, for a current-level node having a subordinate node, after all current-level business operations are completed, a subordinate task is generated and assigned to a subordinate node. The behavior of accepting the superior task by the subordinate node and the behavior of dispatching the subordinate task by the superior node occur simultaneously, and the superior node and the subordinate node cooperate to complete task dispatching and accepting.
S103, constructing a node level business processing model based on a deep neural network, and obtaining the total time consumption of the level node for completing a superior task by using the node level business processing model;
as an implementable embodiment, as shown in fig. 2, the neural network structure adopted by the node level business processing model includes a first long-term memory neural network, a second long-term memory neural network and a deep fully-connected neural network;
correspondingly, the training process of the node current-level business processing model specifically includes:
step A1: cleaning the service data of the nodes to obtain an atomic work vector, a worker vector and the normalization time for completing the current-level service;
specifically, the current-level service is generally divided into a plurality of atomic works according to the assignment of the upper-level tasks, the atomic works can be understood as the minimum working units forming the current-level service, some atomic works can be parallel, and some atomic works have a front-back dependency relationship; each atom work has certain characteristics (such as workload, whether the work depends on the work or not and environment elements), so that the atom work can be represented and described by adopting a characteristic vector, and the atom work vector is obtained by characterizing and describing the atom work through the characteristic vector; similarly, it is assumed that a service activity of a node j at a certain level is provided with m staff members, and each staff member has certain characteristics (for example, whether the staff member is new or old, tired, and can adapt to a parallel task or not), so that a characteristic vector can be used for representing and describing, and the staff member vector is obtained by representing and describing the staff members participating in the service at the level through the characteristic vector; one atomic job corresponds to one atomic job vector and one worker corresponds to one worker vector; by normalized time is meant the real number between the neural network outputs 0 to 1.
Step A2: respectively taking an atomic working vector and a working personnel vector as the input of a first long-short time memory neural network and a second long-short time memory neural network, cascading the output of the last unit of the first long-short time memory neural network and the output of the last unit of the second long-short time memory neural network, inputting the cascading vector into a deep fully-connected neural network, taking the normalization time of the completion of the current-level service as the output of the deep fully-connected neural network, and then adopting supervised learning neural network parameters or simulating the learning neural network parameters through expert data, wherein the trained neural network is the node current-level service processing model.
Assuming that m workers cooperate to complete the current-level business activity, the time cost is determined by each atomic work vector, the worker vector carrying the work, the flow of the atomic work room, and other factors. Since the specific business execution process is often not modelable, deep neural networks are used to model the time cost of the entire level of business.
As shown in fig. 2, the using process of the node present-level service processing model is as follows:
step B1: dividing the service data of the node into a plurality of atom works, and performing representation description on each atom work by using a characteristic vector to obtain an atom work vector of each atom work;
step B2: acquiring all worker information allocated to the node at the current level, and performing representation description on each worker of the node at the current level by adopting the characteristic vector to obtain a worker vector of each worker;
step B3: the method comprises the steps of taking an atomic work vector of each atomic work as the input of each LSTM unit (such as a plurality of LSTM1 units in the figure 2) in a first long-short-term memory neural network, taking a worker vector of each worker as the input of each LSTM unit (such as a plurality of LSTM2 units in the figure 2) in a second long-short-term memory neural network, cascading the output of the last unit of the first long-short-term memory neural network and the output of the last unit of the second long-short-term memory neural network, inputting the cascading vectors into a deep fully-connected neural network (in the figure 2), wherein the deep fully-connected neural network comprises 2 fully-connected layers, each layer comprises 2048 neurons), and outputting the normalization time of the completion of the current-level service by the deep fully-connected neural network (the last fully-connected layer in the figure 2).
As an implementable manner, obtaining the total time consumed by the current-level node to complete the superior task by using the current-level service processing model of the node specifically includes the following steps:
step C1: according to the normalized time output by the node current-level service processing model, the time actually spent by current-level service processing of the current-level node j is calculated according to the formula (1)
Wherein, TmaxRepresents the maximum completion time, T, of the current level of serviceminThe minimum completion time of the current-level service is represented, and y is the normalized time;
step C2: checking whether the current node j has a lower node: if so, the time actually spent by the current-level service processing of each lower-level node is calculated (by step S104), and the time when the current-level node j depends on the lower-level node is calculated according to the formula (2)
If not, then order
Wherein,
representing the time actually spent by the current-level service processing of the lower-level node i of the current-level node j;
step C3: calculating according to a formula (3) to obtain the total time consumption of the current-level node j for completing the superior task
S104, constructing a service model of upper and lower level interaction, and obtaining the total time consumption of the lower level node for completing the upper level task by using the service model of the upper and lower level interaction;
specifically, the interaction refers to a process that a superior assigns a task to a subordinate, and the subordinate receives the superior task and completes the superior task. The service model of the upper and lower level interaction specifically comprises:
step D1: determining a task allocation mode so as to calculate task interaction time under the specific task allocation mode; the task allocation mode comprises an online interactive allocation task mode and an offline interactive allocation task mode;
step D2: according to the task interaction time and the total time consumption of the subordinate node i to finish the superior task
Calculated according to the formula (4) to obtain
Wherein,
indicating a task interaction time required from the current node j to inform its lower node i of the task ready to receive until the lower node i receives the task.
Specifically, a local node first obtains its own direct lower node set, each lower node corresponds to one lower Agent (for example, lower node i corresponds to ith lower Agent), the Agent corresponding to the node is called as a superior Agent, the superior Agent informs its ith lower Agent to prepare to receive a task, and the time required for the superior Agent to receive the task is recorded as the time required for the ith lower node to receive the task
Then, the ith lower Agent recursively calls step S103 to obtain
As an implementation manner, if the task assigning mode adopts on-line interactive task distribution,
and calculating the mean value according to the receiving sequence notified by the node at the current level and the number of the subordinate nodes under the node at the current level, and calculating the variance according to the working state of the subordinate nodes and the stability of the communication link.
For example, assuming that the number of nodes of the immediate lower node of the present level node is H,
w is a receiving sequence number, and T is the maximum receiving time of H subordinate nodes.
As an implementation mode, if the task allocation mode adopts the off-line interactive task allocation mode, the formula (5) is adopted to calculate
Wherein, Ti,goto、Ti,receiveAnd Ti,gobackAnd respectively obtaining the task time and the route time for the subordinate node i to return to the position of the subordinate node according to the notice after the subordinate node i arrives.
In particular, Ti,gotoAnd Ti,gobackGIS map simulation can be introduced by using simulation software package analog. Obtaining task time T after arrivali,receiveAnd can be obtained by a queuing model simulation.
S105, setting configuration parameters of each node according to multi-level complex service conditions, and initializing a simulation setting scene;
specifically, the configuration parameters include an atomic work vector, a staff vector, and upper and lower level interaction parameters of the node. For example, when tasks are distributed using online interactions, the context interaction parameters include mean and variance parameters of the previous and next level online interactions. When the offline interaction distribution task is adopted, the up-down interaction parameters comprise data such as lower-level geographical positions of offline interaction, traffic ways of coming and going, and waiting parameters after arriving.
S106, calling the node current-level service processing model and the upper and lower-level interactive service model for simulation according to given configuration parameters, and recording time data of each node completing service activities in the simulation process;
specifically, simulation and emulation are started from a top-level node, the required current-level business time is simulated by using a neural network model, after the current level is finished, business activity behaviors interacting with a lower level are simulated, time for executing each level of business activity and borne risks are obtained through simulation, and finally, total execution time and total risks are obtained, so that the capacity of identifying and reducing the risks is enhanced.
S107, adjusting configuration parameters of each node, and executing the step S106 again;
specifically, after the atomic working conditions and the staff allocation conditions of each node at each stage are changed, analog simulation is performed again.
And S108, comparing simulation time data under different configuration parameters after the simulation process is finished to obtain the optimal configuration parameters.
For example, in the case where the total number of workers is the same, the merits of the respective equipment plans are compared.
Example 2
As shown in fig. 3, an embodiment of the present invention provides a multi-level complex service intelligent simulation system, including: the system comprises a node level business processing neural network training module, a business level description module, a planning configuration module, a node level business processing simulation module, a node and lower level interactive simulation module and a simulation progress management module;
the node level business processing neural network training module is used for cleaning the business data of the node to obtain an atomic work vector, a worker vector and the normalization time for completion of the level business; respectively taking an atomic working vector and a working personnel vector as the input of a first long-short time memory neural network and a second long-short time memory neural network, cascading the output of the last unit of the first long-short time memory neural network and the output of the last unit of the second long-short time memory neural network, inputting the cascading vector into a deep fully-connected neural network, taking the normalization time of the completion of the current-level service as the output of the deep fully-connected neural network, and then adopting supervised learning neural network parameters or simulating the learning neural network parameters through expert data, wherein the trained neural network is the node current-level service processing model.
The service level description module provides a human-computer interface, and is used for describing the level relation of the complex service by taking each service entity in the complex service as a node so as to determine the superior-inferior relation of each node in each level; the system comprises a plurality of nodes, a task processing module and a task processing module, wherein the nodes are used for determining the node attribute of each node according to the superior-inferior relation of each node to obtain the task content composition of each node;
the planning configuration module provides a human-computer interface and is used for planning configuration parameters of each node according to multi-level complex service conditions; and adjusting configuration parameters of each node;
specifically, the configuration parameters of each node include an atomic work vector, a staff vector, and upper and lower level interaction parameters of the node. For example, when tasks are distributed using online interactions, the context interaction parameters include mean and variance parameters of the previous and next level online interactions. When the offline interaction distribution task is adopted, the up-down interaction parameters comprise data such as lower-level geographical positions of offline interaction, traffic ways of coming and going, and waiting parameters after arriving.
The node local-level service processing simulation module is used for calling a node local-level service processing model to simulate the local-level service processing process according to the given configuration parameters and obtaining the total consumption time of the local-level node for completing the superior task;
specifically, each atomic work vector and each staff vector of the node determined by the imagination configuration module are used as the input of the node level business processing model, so that the node level business simulation is carried out.
The node and lower interactive simulation module is used for calling a service model of upper and lower interaction to simulate the upper and lower service interaction process according to given configuration parameters and obtaining the total time consumption of the lower node for completing the upper task;
specifically, the online interaction parameter data of the node determined by the default configuration module is used for simulating the online interaction process and the interaction time, or the offline interaction data of the node determined by the default configuration module is used for simulating the offline interaction process and the interaction time, and the local service simulation module of the lower node of the node and the lower interaction simulation module is further called by depth-first recursion to obtain the simulation execution time of the lower service.
The simulation progress management module provides a human-computer interface for starting, pausing or stopping a simulation process and can also control the granularity of simulation time.
It should be noted that the multi-level complex service intelligent simulation system provided in the embodiment of the present invention is an embodiment for implementing the method, and specific reference may be made to the method embodiment for functions thereof, which is not described herein again.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. The multi-level complex service intelligent simulation method is characterized by comprising the following steps:
step 1, taking each service entity in the complex service as a node, and describing the hierarchical relationship of the complex service to determine the superior-inferior relationship of each node in each hierarchy;
step 2, determining the node attribute of each node according to the superior-inferior relation of each node to obtain the task content composition of each node;
step 3, constructing a node level business processing model based on a deep neural network, and obtaining the total time consumption of the node level to finish a superior task by utilizing the node level business processing model;
step 4, constructing a service model of the upper-level and lower-level interaction, and obtaining the total time consumption of the lower-level node for completing the upper-level task by using the service model of the upper-level and lower-level interaction;
step 5, setting configuration parameters of each node according to multi-level complex service conditions;
step 6, calling the node current-level service processing model and the upper and lower-level interactive service model for simulation according to the given configuration parameters, and recording time data of each node completing service activities in the simulation process;
step 7, adjusting configuration parameters of each node, and executing step 6 again;
and 8, comparing simulation time data under different configuration parameters after the simulation process is finished to obtain the optimal configuration parameters.
2. The multi-level complex service intelligent simulation method according to claim 1, wherein in step 2, the node attributes comprise a top level node, a middle level node and a bottom level node; in a corresponding manner, the first and second electrodes are,
the task content constitution of the top level node comprises the following steps: processing the current-level service and assigning lower-level tasks;
the task content constitution of the middle level node comprises the following steps: receiving a superior task, processing a current-level service and dispatching a subordinate task;
the task content of the bottommost node comprises the following steps: receiving the superior task and processing the current service.
3. The multi-level complex business intelligent simulation method according to claim 1, wherein in step 3, the neural network structure adopted by the node level business processing model comprises a first long-term and short-term memory neural network, a second long-term and short-term memory neural network and a deep full-link neural network;
the training process of the node current-level business processing model specifically comprises the following steps:
cleaning the service data of the nodes to obtain an atomic work vector, a worker vector and the normalization time for completing the current-level service; the atomic work is the minimum work unit forming the current-level service, the atomic work vector is obtained by performing representation description on the atomic work through the feature vector, and the staff vector is obtained by performing representation description on staff participating in the current-level service through the feature vector; one atomic job corresponds to one atomic job vector and one worker corresponds to one worker vector;
respectively taking an atomic working vector and a working personnel vector as the input of a first long-short time memory neural network and a second long-short time memory neural network, cascading the output of the last unit of the first long-short time memory neural network and the output of the last unit of the second long-short time memory neural network, inputting the cascading vector into a deep fully-connected neural network, taking the normalization time of the completion of the current-level service as the output of the deep fully-connected neural network, and then adopting supervised learning neural network parameters or simulating the learning neural network parameters through expert data, wherein the trained neural network is the node current-level service processing model.
4. The multi-level complex service intelligent simulation method according to claim 3, wherein the step 3 specifically further comprises:
according to the normalized time output by the node current-level service processing model, the time actually spent by current-level service processing of the current-level node j is calculated according to the formula (1)
Wherein, TmaxRepresents the maximum completion time, T, of the current level of serviceminThe minimum completion time of the current-level service is represented, and y is the normalized time;
checking whether the current node j has a lower node: if the current service processing time exists, the time actually spent by the current service processing of each lower node is calculated,and calculating the time of the node j depending on the lower node according to the formula (2)
If not, then order
Wherein,
the time actually spent by the processing of the current-level service of the lower-level node i of the current-level node j is represented;
calculating according to a formula (3) to obtain the total time consumption of the current-level node j for completing the superior task
5. The multi-level complex service intelligent simulation method according to claim 4, wherein the service model of the upper and lower level interaction specifically comprises:
determining a task allocation mode so as to calculate task interaction time under the specific task allocation mode; the task allocation mode comprises an online interactive allocation task mode and an offline interactive allocation task mode;
according to the task interaction time and the total time consumption of the subordinate node i to finish the superior task
Calculated according to the formula (4)
Wherein,
indicating a task interaction time required from the current node j to inform its lower node i of the task ready to receive until the lower node i receives the task.
6. The multi-level complex business intelligence simulation method of claim 5, wherein if on-line interactive task allocation is adopted,
and calculating the mean value according to the receiving sequence notified by the node at the current level and the number of the subordinate nodes under the node at the current level, and calculating the variance according to the working state of the subordinate nodes and the stability of the communication link.
7. The multi-level complex service intelligent simulation method of claim 5, wherein if the tasks are distributed by offline interaction, the tasks are calculated by formula (5)
Wherein, Ti,goto、Ti,receiveAnd Ti,gobackAnd respectively obtaining the task time and the route time for the subordinate node i to return to the position of the subordinate node according to the notice after the subordinate node i arrives.
8. The multi-hierarchy intelligent simulation method for complex services according to claim 3, wherein in step 5, the configuration parameters include an atomic work vector, a staff vector and upper and lower level interaction parameters of a node.
9. The multi-level complex service intelligent simulation system is characterized by comprising: the system comprises a service level description module, a scenario configuration module, a node home-level service processing simulation module, a node and lower-level interaction simulation module and a simulation progress management module;
the service level description module is used for describing the level relation of the complex service by taking each service entity in the complex service as a node so as to determine the superior-inferior relation of each node in each level; the system comprises a plurality of nodes, a task processing module and a task processing module, wherein the nodes are used for determining the node attribute of each node according to the superior-inferior relation of each node to obtain the task content composition of each node;
the planning configuration module is used for planning configuration parameters of each node according to multi-level complex service conditions; and adjusting configuration parameters of each node;
the node local-level service processing simulation module is used for calling a node local-level service processing model to simulate a local-level service processing process according to given configuration parameters and obtaining the total time consumed by the local-level node to finish a superior task;
the node and lower interactive simulation module is used for calling a service model of upper and lower interaction to simulate the upper and lower service interaction process according to given configuration parameters and obtaining the total time consumption of the lower node for completing the upper task;
the simulation progress management module is used for starting, pausing or stopping a simulation process.
10. The multi-level complex business intelligence simulation system of claim 9, further comprising: a node current-level service processing neural network training module;
the node level business processing neural network training module is used for cleaning the business data of the node to obtain an atomic work vector, a worker vector and the normalization time for completion of the level business; respectively taking an atomic working vector and a working personnel vector as the input of a first long-short time memory neural network and a second long-short time memory neural network, cascading the output of the last unit of the first long-short time memory neural network and the output of the last unit of the second long-short time memory neural network, inputting the cascade vector into a deep fully-connected neural network, taking the normalized time of the completion of the current-level service as the output of the deep fully-connected neural network, and then adopting supervised learning to train neural network parameters or simulating the learning to train the neural network parameters through expert data, wherein the trained neural network is the node current-level service processing model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210269296.1A CN114611401A (en) | 2022-03-18 | 2022-03-18 | Multi-level complex service intelligent simulation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210269296.1A CN114611401A (en) | 2022-03-18 | 2022-03-18 | Multi-level complex service intelligent simulation method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114611401A true CN114611401A (en) | 2022-06-10 |
Family
ID=81864779
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210269296.1A Pending CN114611401A (en) | 2022-03-18 | 2022-03-18 | Multi-level complex service intelligent simulation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114611401A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116011257A (en) * | 2023-03-07 | 2023-04-25 | 北京创奇视界科技有限公司 | Flow simulation method and device for multi-level task |
-
2022
- 2022-03-18 CN CN202210269296.1A patent/CN114611401A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116011257A (en) * | 2023-03-07 | 2023-04-25 | 北京创奇视界科技有限公司 | Flow simulation method and device for multi-level task |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ashouraie et al. | Priority-based task scheduling on heterogeneous resources in the Expert Cloud | |
| Levitt et al. | The virtual design team: Simulating how organization structure and information processing tools affect team performance | |
| Chang et al. | Time-line based model for software project scheduling with genetic algorithms | |
| Karagiannis | BPMS: business process management systems | |
| Suri et al. | Comparative analysis of software effort estimation techniques | |
| US20150161539A1 (en) | Decision support system for project managers and associated method | |
| Zhu et al. | Towards a new paradigm for management of complex engineering projects: A system-of-systems framework | |
| Lopez et al. | Distributed reinforcement learning in emergency response simulation | |
| CN113112107A (en) | Transfer planning method, device, server and storage medium | |
| Zhang et al. | Modelling and simulation of the task scheduling behavior in collaborative product development process | |
| CN114611401A (en) | Multi-level complex service intelligent simulation method and system | |
| Hashemipour et al. | A disaster multiagent coordination simulation system to evaluate the design of a first‐response team | |
| CN109901820A (en) | A kind of optimization method for the Airborne Software agile development process meeting DO-178B/C | |
| CN112508201A (en) | Acitivti 7.0-based artificial intelligence general training method and system | |
| Pawlewski et al. | Application potential of Agent Based Simulation and Discrete Event Simulation in enterprise integration modelling concepts | |
| Manousakis et al. | A Practical Approach for Resource-Constrained Project Scheduling | |
| Cao et al. | Online cost-rejection rate scheduling for resource requests in hybrid clouds | |
| Fu et al. | Robust partial order schedules for rcpsp/max with durational uncertainty | |
| CN109657958B (en) | Modeling method of digital information system and digital information system | |
| Voudouris et al. | ARMS: An automated resource management system for British Telecommunications plc | |
| Sgouros et al. | Creating web-based, meta-simulation environments for social dynamics in an interactive framework for public policy analysis and design | |
| Mutingi | System dynamics of manpower planning strategies under various demand scenarios | |
| Peck et al. | Systems dynamic modeling: planning beyond the worker | |
| CN118428697B (en) | Scheduling method, system, equipment and storage medium for mine equipment overhaul resources | |
| Mihret et al. | Simulation-based recommendation generation for heterogeneous systems participating in a collaborative work |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: 450000 Science Avenue 62, Zhengzhou High-tech Zone, Henan Province Applicant after: Information Engineering University of the Chinese People's Liberation Army Cyberspace Force Address before: No. 62 Science Avenue, High tech Zone, Zhengzhou City, Henan Province Applicant before: Information Engineering University of Strategic Support Force,PLA Country or region before: China |