CN-122022231-A - Dynamic allocation method and device for engineering resources

CN122022231ACN 122022231 ACN122022231 ACN 122022231ACN-122022231-A

Abstract

The invention relates to the technical field of engineering digital management and intelligent scheduling, in particular to a dynamic allocation method and device of engineering resources. The method comprises the steps of obtaining multi-source heterogeneous physical data of an engineering full life cycle, preprocessing the multi-source heterogeneous physical data, utilizing gradients to promote nonlinear relations between physical parameters of an excavation environment and resource consumption of a decision tree based on the preprocessed physical data, predicting physical resource demand correction coefficients and working condition risk levels of all construction nodes, and adopting a river horse optimization algorithm, the physical resource demand correction coefficients and the working condition risk levels to output optimal resource allocation vectors on the premise of meeting physical constraints with the aim of minimizing physical supply and demand deviation, system operation energy consumption and safety out-of-limit penalty. The method and the device can solve the problems that static planning in the prior art is difficult to match with real-time changing field requirements, and a general algorithm is difficult to process complex physical constraints, so that resource allocation is uneven.

Inventors

LIU HONGZHI
TU QINGBO
SHI LICHI
Cao Sixuan
WANG YONGLI
KANG FANG
GAO HE
ZHANG CAN
ZHANG HONGYANG
YANG XIAOYING
WANG YANMEI
LI KAI
SONG JIANBO

Assignees

国网山东省电力公司经济技术研究院
华北电力大学

Dates

Publication Date: 20260512
Application Date: 20251212

Claims (10)

1. A method for dynamically allocating engineering resources, comprising: acquiring multi-source heterogeneous physical data of an engineering full life cycle, and preprocessing the multi-source heterogeneous physical data; Based on the preprocessed physical data, utilizing a gradient to promote a nonlinear relation between physical parameters of the decision tree mining environment and resource consumption, and predicting physical resource demand correction coefficients and working condition risk levels of all construction nodes; And (3) adopting a river horse optimization algorithm, the physical resource demand correction coefficient and the working condition risk level, and outputting an optimal resource allocation vector on the premise of meeting physical constraint by taking the minimum physical supply and demand deviation, the system operation energy consumption and the safety out-of-limit penalty as targets.
2. The dynamic allocation method of engineering resources according to claim 1, wherein predicting the physical resource demand correction coefficient and the working condition risk level of each construction node based on the preprocessed physical data by utilizing a gradient to promote a nonlinear relation between physical parameters of a decision tree mining environment and resource consumption comprises: based on the preprocessed physical data, extracting multidimensional working condition characteristics, and constructing a high-dimensional characteristic vector according to the multidimensional working condition characteristics; Defining a prediction target, and setting a physical resource demand correction coefficient as a core target of the prediction target; based on the physical resource demand correction coefficient, constructing a physical adaptive asymmetric loss function, so that a penalty weight higher than resource surplus is applied to a prediction deviation corresponding to resource shortage; Training a gradient lifting decision tree by taking the high-dimensional physical feature vector as input, the physical resource demand correction coefficient as output and the minimized physical adaptive asymmetric loss function as a criterion; and inputting the pre-processed physical data updated in real time into the gradient lifting decision tree after training is completed, and outputting the physical resource demand correction coefficient and the working condition risk level of each construction node.
3. The method for dynamically allocating engineering resources according to claim 2, wherein the physical adaptive asymmetric loss function is: ; Wherein, the Representing a physically adapted asymmetric loss function, A first weight is indicated and a second weight is indicated, Represents a second weight, an , Representing the actual value of the physical resource demand modification factor, And representing the predicted value of the physical resource demand correction coefficient.
4. The method for dynamically allocating engineering resources according to claim 1, wherein the adopting the river horse optimization algorithm, the physical resource demand correction coefficient and the working condition risk level, with the objective of minimizing physical supply and demand deviation, system operation energy consumption and safety out-of-limit penalty, outputs an optimal resource allocation vector on the premise of meeting physical constraints, comprises: Constructing a mathematical model based on the preprocessed physical data, the physical resource demand correction coefficient and the working condition risk level, wherein the mathematical model comprises decision variables, constraint conditions and multi-objective fitness functions; generating an initial resource allocation scheme based on the mathematical model to form an initial population of a river horse optimization algorithm; Dynamically adjusting the collar-to-ground weight and the environmental disturbance factor in the river horse optimization algorithm based on the population individual fitness performance and the preprocessed real-time physical environment data; Combining three behaviors of collar conservation, group migration and water level adaptation, updating the positions of individuals in the group based on the adjusted collar weight and the environment disturbance factor, and correcting and ensuring that the new positions meet the constraint conditions; Judging whether to terminate the iteration, if yes, outputting a global optimal resource allocation vector, if not, returning to the step of dynamically adjusting the collar-to-ground weight and the environmental disturbance factor in the river horse optimization algorithm, readjusting parameters, and continuing the iteration.
5. The method for dynamically allocating engineering resources according to claim 4, wherein the decision variables are: , , Wherein, the method comprises the steps of, A set of decision variables is represented and, Represent the first Class standard shift equivalent assignment to the first The number of the construction nodes to be constructed, Represents the number of resource types and, Representing the number of construction nodes; the constraint conditions comprise total resource constraint, physical security constraint and dynamic demand constraint; the total resource constraint is: Wherein, the method comprises the steps of, Representing the total resources; The physical security constraints are: Wherein, the method comprises the steps of, Representing the real-time safety margin of the node, Representing a minimum safety threshold; The dynamic demand constraint is that when the deviation rate of a certain node predicted by the gradient lifting decision tree is larger than a preset value, the dynamic demand constraint is that according to the following condition Updating the upper resource limit of the node, wherein, Representing the updated upper limit of node resources, Represents the node resource upper limit before the update, The correction coefficient is represented by a number of coefficients, Representing a historical prediction bias rate of the gradient lifting decision tree; The multi-objective fitness function is: ; Wherein, the Representing a multi-objective fitness function, 、、 Respectively represent the importance coefficients of the characteristics of the gradient-lifting decision tree mining, The matching degree of the supply and the demand is expressed, The physical energy consumption is represented by the number of the physical energy consumption, A security penalty is indicated and a security penalty is indicated, Representing the predicted need for a gradient-lifting decision tree, Represent the first The unit energy consumption similar to the standard table class equivalent, Represent the first Class standard table class equivalent to th Scheduling distance of individual construction nodes.
6. The method for dynamically allocating engineering resources according to claim 5, wherein an initialization formula of the initial population is: ; Wherein, the Represent the first Of individuals in the individual population, the first Class standard shift equivalent assignment to the first The number of the construction nodes to be constructed, , Indicating the size of the population, Represent the first The upper limit of node resources of each construction node, Representing the uniform random number of the interval [ a, b ], The optimal utilization rate is indicated to be high, Represent the first The historical average resource occupancy of each construction node, Representing the historical average resource occupancy of all construction nodes, , Representing the number of historical cycles.
7. The method for dynamically allocating engineering resources according to claim 6, wherein dynamically adjusting the collar-to-land weight and the environmental disturbance factor in the river horse optimization algorithm based on the population individual fitness performance and the preprocessed real-time physical environment data comprises: According to Dynamically adjusting the collar weight in the river horse optimization algorithm; Wherein, the Representing the weight of the territory, Representing the initial collar weight of the earth, Indicating the selection of the pressure coefficient(s), A fitness value representing the individual is presented, Represents a global optimum fitness value and, Indicating the average fitness of the population, Representing a smoothing term; And determining an environmental disturbance factor based on the preprocessed real-time rainfall and wind speed, wherein the environment is worse, and the environmental disturbance factor is larger.
8. The method of claim 7, wherein updating the locations of the population individuals based on the adjusted territory weights and the environmental perturbation factors comprises: According to Updating the positions of individuals in the population; Wherein, the Represent the first The population individual positions corresponding to the multiple iterations, Represent the first The population individual positions corresponding to the multiple iterations, The local update component is represented and, The global update component is represented and, Representing the component of the environmental disturbance, 、 Random numbers respectively representing the intervals of [0,1], Represent the first Individual at the first The historical optimal position of the dimension is determined, Representing the population at the first The global optimum position of the dimension is determined, The weight coefficients representing the global update are represented, The weight coefficient representing the disturbance is represented by, Representing the factor of the disturbance of the environment, Represent the first The upper limit of the dimensional position, Represent the first Lower limit of dimensional position.
9. The method for dynamically allocating engineering resources according to any one of claims 1 to 8, wherein, after adopting a river horse optimization algorithm, the physical resource demand correction coefficient and the working condition risk level, with the objective of minimizing a physical supply and demand deviation, a system operation energy consumption and a safety out-of-limit penalty, outputting an optimal resource allocation vector on the premise of meeting physical constraints, further comprises: When physical monitoring data is out of limit or the change of the physical resource demand correction coefficient predicted by the gradient lifting decision tree exceeds a threshold value, triggering the river horse optimization algorithm to reallocate resources, and generating a scheduling instruction to be sent to an execution end.
10. A dynamic allocation device for engineering resources, comprising: The acquisition module is used for acquiring multi-source heterogeneous physical data of the engineering full life cycle; the preprocessing module is used for preprocessing the multi-source heterogeneous physical data; The prediction module is used for predicting the physical resource demand correction coefficient and the working condition risk level of each construction node by utilizing the nonlinear relation between the physical parameters of the gradient lifting decision tree mining environment and the resource consumption based on the preprocessed physical data; The processing module is used for outputting an optimal resource allocation vector on the premise of meeting physical constraint by adopting a river horse optimization algorithm, the physical resource demand correction coefficient and the working condition risk level and taking the minimum physical supply and demand deviation, the system operation energy consumption and the safety out-of-limit penalty as targets.

Description

Dynamic allocation method and device for engineering resources Technical Field The invention relates to the technical field of engineering digital management and intelligent scheduling, in particular to a dynamic allocation method and device of engineering resources. Background In large engineering construction such as power grid, traffic, water conservancy and the like, the on-site physical environment is complex and changeable, the dynamic changes of physical parameters such as topography, environmental weather, equipment states and the like directly influence the construction process, and the resource allocation and management and control are required to adapt to the actual working conditions so as to ensure the efficient propulsion of engineering. The current engineering resource allocation is dependent on static design drawings or experience quota, a general optimization algorithm is often used for resource scheduling, a data mining technology is also gradually applied to engineering management, and the aspects of macroscopic progress control or cost estimation and the like are mainly focused. However, the prior art has obvious defects that in actual engineering, physical parameters can change in real time, so that static planning is difficult to match with field requirements of real-time change, and a general algorithm is difficult to process complex physical constraints, so that resource allocation is uneven. Disclosure of Invention The embodiment of the invention provides a dynamic allocation method and a dynamic allocation device for engineering resources, which are used for solving the problems that static planning in the prior art is difficult to match with real-time changing field requirements, and a general algorithm is difficult to process complex physical constraints, so that resource allocation is uneven. In a first aspect, an embodiment of the present invention provides a method for dynamically allocating engineering resources, including: acquiring multi-source heterogeneous physical data of an engineering full life cycle, and preprocessing the multi-source heterogeneous physical data; Based on the preprocessed physical data, utilizing a gradient to promote a nonlinear relation between physical parameters of the decision tree mining environment and resource consumption, and predicting physical resource demand correction coefficients and working condition risk levels of all construction nodes; And (3) adopting a river horse optimization algorithm, the physical resource demand correction coefficient and the working condition risk level, and outputting an optimal resource allocation vector on the premise of meeting physical constraint by taking the minimum physical supply and demand deviation, the system operation energy consumption and the safety out-of-limit penalty as targets. In a second aspect, an embodiment of the present invention provides a dynamic allocation apparatus for engineering resources, including: The acquisition module is used for acquiring multi-source heterogeneous physical data of the engineering full life cycle; the preprocessing module is used for preprocessing the multi-source heterogeneous physical data; The prediction module is used for predicting the physical resource demand correction coefficient and the working condition risk level of each construction node by utilizing the nonlinear relation between the physical parameters of the gradient lifting decision tree mining environment and the resource consumption based on the preprocessed physical data; The processing module is used for outputting an optimal resource allocation vector on the premise of meeting physical constraint by adopting a river horse optimization algorithm, the physical resource demand correction coefficient and the working condition risk level and taking the minimum physical supply and demand deviation, the system operation energy consumption and the safety out-of-limit penalty as targets. In a third aspect, an embodiment of the present invention provides a terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the method for dynamically allocating engineering resources according to the first aspect or any one of the possible implementations of the first aspect, when the processor executes the computer program. In a fourth aspect, embodiments of the present invention provide a computer readable storage medium storing a computer program which, when executed by a processor, implements the steps of the method for dynamic allocation of engineering resources according to the first aspect or any one of the possible implementations of the first aspect. The embodiment of the invention provides a dynamic allocation method and a dynamic allocation device for engineering resources, which ensure that the basis of resource allocation is changed from 'experience estimation' to 'physical data driving' by