A method for optimizing shared node load scheduling parameters

Abstract

The present application belongs to the technical field of shared equipment scheduling control methods, and in particular, relates to a method for optimizing shared node load scheduling parameters. The method includes the following steps: determining the basic data of shared nodes in the system; determining the shared nodes that need to be load scheduled in the shared system; constructing a load transfer optimization model under natural flow; constructing a shared node utilization and current load rate calculation model; constructing a system effective load transfer target; constructing a shared node load scheduling optimization model; solving the shared node load scheduling optimization model to determine the load transfer target and transfer path of each load node. The shared node load scheduling parameter optimization method based on the present application can not only efficiently and simply obtain the necessary parameters required for load scheduling, but also effectively compress the amount of data that needs to be processed and analyzed during the data calculation process, reduce the system prediction and analysis costs, and make the application and implementation of the method more convenient and efficient, with less demand on software and hardware.

Description

Technical Field

[0001] The present application belongs to the technical field of shared device scheduling control methods, and in particular, relates to a method for optimizing shared node load scheduling parameters. Background Art

[0002] With the continuous development of the sharing industry, a large number of shared devices have been put into the market. Since the utilization rate of shared devices is affected by their purpose and various external factors, it is difficult to achieve even distribution of utilization. In most sharing systems, shared nodes used to provide service supply, in addition to serving their own areas, often need to dynamically schedule loads with external shared nodes to cope with periodic usage needs. This load scheduling is generally carried out by shared service providers based on the real-time needs of shared devices. Therefore, there is a lag, which makes it difficult to respond to the increasingly competitive service market. Summary of the invention

[0003] The purpose of the present application is to provide a method for optimizing shared node load scheduling parameters for planning and designing load scheduling targets in advance and optimizing load scheduling data.

[0004] To achieve the above objectives, this application adopts the following technical solutions.

[0005] A method for optimizing shared node load scheduling parameters comprises the following steps:

[0006] Step 1: Determine the basic data of shared nodes in the system

[0007] Specifically, according to the system scheduling range, coordinates are formulated to determine the coordinate positions of all shared nodes and the maximum load capacity of each shared node;

[0008] Step 2: Determine the shared nodes that need to be load dispatched in the shared system;

[0009] For any node j’s load state, the load balancing target within the construction cycle is:

[0010]

[0011] It means that the current load rate of shared node j is too low and the load needs to be adjusted in;

[0012] Indicates that the current load rate of shared node j is too high and needs to be transferred out;

[0013] It indicates that the current load rate of shared node j meets the requirements;

[0014] where f j is the current load rate of the shared node; fjmax is the maximum load rate of the shared node;

[0015] Empirical parameter θ j , (0≤θ j ≤1) is obtained by statistically analyzing the actual operation process of each shared node j for a period of time, which satisfies the load rate of the shared node when When the range is set, the utilization index of the shared nodes can be guaranteed to meet the requirements; based on the load balancing target, the shared nodes that do not meet the balancing target are selected as the shared nodes that need to be loaded;

[0016] Step 3: Constructing a load transfer optimization model under natural flow

[0017] Specifically, the natural load flow of each load node during the statistical period is represented by the load flow probability symbol k ij ∈[0,1] represents the probability of load transfer between different shared nodes, and constructs the optimization objective based on the minimum natural flow path of load between shared nodes. N is the set of all shared nodes that need to transfer load;

[0018] Step 4: Build a calculation model for shared node utilization and current load rate

[0019] Based on the load balancing target within the cycle, the utilization μ of the shared node j is constructed j With the current load factor f j The relationship is:

[0020] Step 5: Build system payload transfer target

[0021] System effective load transfer rate target

[0022] Where R k is the total amount of all shared devices in the load input and output terminal k, J is the total amount of shared nodes in the load input and output terminal k, μ j is the utilization rate of the shared node j within the load input and output terminal k;

[0023] Step 6: Build a shared node load scheduling optimization model

[0024]

[0025] Step 7: Solve the shared node load scheduling optimization model to determine the load transfer target and transfer path of each load node.

[0026] The step 3 of the further improvement or preferred implementation scheme of the above-mentioned shared node load scheduling parameter optimization method also includes clustering simplification of the shared nodes that need to be load scheduled, specifically including:

[0027] 3.1. Determine the load transfer distance between shared nodes, locate the two shared nodes with the largest load transfer distance, and set these two shared nodes as cluster centers;

[0028] 3.2. Eliminate the shared nodes that have been set as cluster centers, and determine the shared nodes that maximize the distance product with the known cluster center among the remaining shared nodes as the new cluster center;

[0029] 3.3. Repeat the above step 3 and select several cluster centers;

[0030] 3.4. After determining the shared node as the cluster center, calculate the distances from the shared nodes of other non-cluster centers to the shared nodes corresponding to each cluster center, and divide all shared nodes of non-cluster centers into a category corresponding to the nearest cluster center according to the standard of minimum load transfer distance;

[0031] 3.5. Calculate the coordinates of all shared nodes in each category, average them and obtain new coordinates as new cluster centers, repeat the above step 5, and re-cluster all shared nodes that are not cluster centers based on the new cluster centers until the maximum number of iterations is reached or the average error meets the standard requirements;

[0032] 3.6. Obtain the final clustering results, and use the coordinates corresponding to each cluster center as the load nodes that ultimately need to be transferred.

[0033] Its beneficial effects are:

[0034] The shared node load scheduling parameter optimization method based on the present application can not only efficiently and simply obtain the necessary parameters required for load scheduling, but also effectively compress the amount of data that needs to be processed and analyzed during the data calculation process. Combined with the set-improved clustering scheme, it can reduce the pressure on the system computing power on the original basis, increase the data acquisition speed, and reduce the system prediction and analysis costs, making the application and implementation of this method more convenient and efficient, and requiring less software and hardware. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is the relationship curve between the utilization rate of shared node j and the current load rate. DETAILED DESCRIPTION

[0036] For any shared system, the distance between its nodes directly affects the endurance of the supporting shared equipment, the size of the service range, and the stability of the service utilization. In most cases, when the service area and the number of shared nodes V in the area are known, the distance d between shared nodes i and shared nodes j is ij The smaller the requirement, the better. There may be multiple load transfer paths between multiple shared nodes in the system. Therefore, based on the above, first determine the one with the shortest distance among the multiple load transfer paths. For several shared nodes that need to be load dispatched, in order to minimize the cost of the system when performing load dispatch, it is necessary to ensure that the dispatching distance required when executing the load dispatch is minimized. This is essentially a TSP problem. Based on this, the main steps of the shared node load dispatching parameter optimization method of this application are as follows:

[0037] Step 1: Determine the basic data of shared nodes in the system

[0038] The basic data include the coordinate positions of the shared nodes based on the area where the system is located, the available load transfer paths;

[0039] Step 2: determine n shared nodes that need to be load dispatched in the shared system;

[0040] In order to ensure that the load rate of each contributing node in the region is as high as possible, for the load state of any node j, although full load can ensure that the load rate of the current shared node at a specific time point reaches the highest, due to the transfer of loads between shared nodes, when a shared node is in a full load state, the load transferred in by the previous shared node will not be able to be quickly accessed, and at the same time, it will have an adverse effect on the load rate of the next shared node; therefore, for the entire system, the optimal load rate of each shared node is not the maximum load rate. Generally, when the real-time load is the maximum load f jmax When the number of nodes is half of the original number, the service utilization rate of the shared nodes can be maximized;

[0041] But in fact, since each shared node in the shared system is in a dynamic adjustment process, for the entire system, the optimal load rate f of any single shared node j is j It not only involves the maximum load f of the node jmax , considering the optimal load rate f of shared node j j The continuous access and transfer of loads between adjacent shared nodes is actually a dynamic adjustment process. Therefore, the empirical parameter θ can be obtained through data statistics. j , (0≤θ j ≤1); therefore, the load balancing target within the cycle can be expressed as: when Indicates that the current load rate of shared node j is too small. In order to improve the system utilization rate, the load needs to be adjusted. Indicates that the current load rate of shared node j is large and needs to be transferred out; when It indicates that the current load rate of shared node j meets the requirements;

[0042] The empirical parameter θ j , (0≤θ j ≤1) is obtained by statistically analyzing the actual operation process of each shared node j for a period of time. When the load rate of the shared node is When the scope is met, it can ensure that the technical indicators such as the utilization rate of shared nodes meet the requirements;

[0043] For large-scale shared systems, due to the large number of shared nodes, a large amount of data calculation and processing work may be required, affecting the data processing efficiency. Considering that a large number of shared devices placed in the same shared node are closely related to the customer's activity range when in use, the circulation and use of shared devices presents a clustering characteristic, and at the same time affects the load flow characteristics within each shared node. This is reflected in the probability of load transfer between two shared nodes with a long load transfer distance being much lower than the probability of load transfer between two load nodes with a small distance. Therefore, in order to obtain the optimal solution of the aforementioned optimization objective function, the shared nodes that need to be load dispatched can be clustered. By considering the clustered flow of shared devices in specific shared nodes, the rationality of the solution can be optimized while reducing the scale of data processing; specifically, the following clustering scheme can be adopted:

[0044] 1. Determine the load transfer distance among the aforementioned n shared nodes, and determine the two shared nodes with the largest load transfer distance; set these two shared nodes as cluster centers;

[0045] 2. Eliminate the shared nodes that have been set as cluster centers, and determine the shared nodes that maximize the distance product with the known cluster center, that is, max(d1×d2×d3....) among the remaining shared nodes, as the new cluster center; d i It refers to the distance from the center of each cluster i;

[0046] 3. Repeat the above step 3 to select several cluster centers;

[0047] 4. After determining the shared node as the cluster center, calculate the distance from the shared nodes of other non-cluster centers to the shared nodes corresponding to each cluster center, and divide all shared nodes of non-cluster centers into a category corresponding to their nearest cluster center according to the standard of minimum load transfer distance;

[0048] 5. Calculate the coordinates of all shared nodes in each category, average them and obtain new coordinates as new cluster centers, repeat the above step 5, and re-cluster all shared nodes that are not cluster centers based on the new cluster centers until the maximum number of iterations is reached or the average error meets the standard requirements;

[0049] In actual operation, there is not a natural flow of load between any two shared nodes, but a selective and directional flow based on the actual situation. Assuming that the load flow probability mark k ij ∈[0,1] means, k ij =1 indicates that there is a transfer from shared node i to shared node j, k ij = 0 indicates that there is no transfer from shared node i to shared node j, then the optimization goal based on the minimum load scheduling path between shared nodes can be obtained

[0050] Taking the aforementioned range load rate as the standard, we can draw a curve of the relationship between the utilization rate of shared node j and the current load rate, as shown in Figure 1 As shown, the utilization rate of shared node j can be expressed as

[0051]

[0052] In a shared system, due to the homogeneity of shared devices, the transfer of two shared devices with different IDs to and from the same shared node j in a short period of time can be regarded as the continuous transfer of a single shared device and the interaction between the shared device and the shared node j can be ignored. j =1, the corresponding shared nodes can be regarded as balanced and no load adjustment is required. When analyzing the optimal solution for system load scheduling, only μ j ≠1 corresponding to the shared nodes;

[0053] Based on the above steps, several cluster centers are obtained. The distribution and characteristics of these cluster centers directly affect the results and efficiency of the system when transferring load. On this basis, all shared nodes in each cluster can be fuzzyized into a load input and output terminal. The input-output relationship of each load input and output terminal is internally self-consistent and only related to other external load input and output terminals. For the entire shared system, the average system load rate of these load input and output terminals can be used to simplify the calculation of the effective utilization rate of the shared system. The load transfer operation often only involves the location transfer of the shared equipment, and the load transfer operation needs to take into account the availability of the shared equipment and the re-planning of the service plan. In a stable shared system, for the K load input and output terminals that need to be load-scheduled, the average load rate should be maximized, that is, the system effective load transfer rate target

[0054] Where R k is the total amount of all shared devices in the load input and output terminal k, J is the total amount of shared nodes in the load input and output terminal k, μ j is the utilization rate of the shared node j within the load input and output terminal k;

[0055] Based on the above content, the optimization target model for load scheduling can be constructed:

[0056]

[0057] Based on the above optimization target model, the optimal load transfer parameters can be determined, that is, the utilization rate μ of each load node after load transfer j , and based on μ j The target load rate is obtained by reverse calculation, and load dispatch can be performed based on the difference between the load rate target and the previous load rate.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application, rather than to limit the scope of protection of the present application. Although the present application has been described in detail with reference to the preferred embodiments, ordinary technicians in this field should understand that the technical solution of the present application can be modified or replaced by equivalents without departing from the essence and scope of the technical solution of the present application.

Claims

1. A method for optimizing shared node load scheduling parameters, characterized in that: The steps include: Step 1: Determine the basic data of shared nodes in the system Specifically, according to the system scheduling range, coordinates are formulated to determine the coordinate positions of all shared nodes and the maximum load capacity of each shared node; Step 2: Determine the shared nodes that need to be load dispatched in the shared system; For any node j’s load state, the load balancing target within the construction cycle is: It means that the current load rate of shared node j is too low and the load needs to be adjusted in; Indicates that the current load rate of shared node j is too high and needs to be transferred out; It indicates that the current load rate of shared node j meets the requirements; where f j is the current load rate of the shared node; f jmax is the maximum load rate of the shared node; Empirical parameter θ j , (0≤θ j ≤1) is obtained by statistically analyzing the actual operation process of each shared node j for a period of time, which satisfies the load rate of the shared node when When the range is set, the utilization index of the shared nodes can be guaranteed to meet the requirements; based on the load balancing target, the shared nodes that do not meet the balancing target are selected as the shared nodes that need to be loaded; Step 3: Constructing a load transfer optimization model under natural flow Specifically, the natural load flow of each load node during the statistical period is represented by the load flow probability symbol k ij ∈[0,1] represents the probability of load transfer between different shared nodes, and constructs the optimization objective based on the minimum natural flow path of load between shared nodes. N is the set of all shared nodes that need to transfer load; Step 4: Build a calculation model for shared node utilization and current load rate Based on the load balancing target within the cycle, the utilization μ of the shared node j is constructed j With the current load factor f j The relationship is: Step 5: Build system payload transfer target System effective load transfer rate target Where R k is the total amount of all shared devices in the load input and output terminal k, J is the total amount of shared nodes in the load input and output terminal k, μ j is the utilization rate of the shared node j within the load input and output terminal k; Step 6: Construct a shared node load scheduling optimization model Step 7: Solve the shared node load scheduling optimization model to determine the load transfer target and transfer path of each load node.

2. A method for optimizing shared node load scheduling parameters according to claim 1, characterized in that: The step three also includes clustering and simplifying the shared nodes that need to be load dispatched, specifically including: 3.

1. Determine the load transfer distance between shared nodes, locate the two shared nodes with the largest load transfer distance, and set these two shared nodes as cluster centers.

3. A method for optimizing shared node load scheduling parameters according to claim 2, characterized in that: The clustering simplification of the shared nodes that need to be load dispatched also includes: 3.

2. Eliminate the shared nodes that have been set as cluster centers, and determine the shared nodes that maximize the distance product with the known cluster center among the remaining shared nodes as the new cluster center.

4. A method for optimizing shared node load scheduling parameters according to claim 3, characterized in that: The clustering simplification of the shared nodes that need to be load dispatched also includes: 3.

3. Repeat the above step 3.1 and select several cluster centers; 3.

4. After determining the shared node as the cluster center, calculate the distance from the shared nodes of other non-cluster centers to the shared nodes corresponding to each cluster center, and divide all shared nodes of non-cluster centers into a category corresponding to their nearest cluster center according to the standard of minimum load transfer distance.

5. A method for optimizing shared node load scheduling parameters according to claim 4, characterized in that: The clustering simplification of the shared nodes that need to be load dispatched also includes: 3.

5. Calculate the coordinates of all shared nodes in each category, average them and obtain new coordinates as new cluster centers. Repeat the above step 3.4, and re-cluster all shared nodes that are not cluster centers based on the new cluster centers until the maximum number of iterations is reached or the average error meets the standard requirements. 3.

6. Obtain the final clustering results, and use the coordinates corresponding to each cluster center as the load nodes that ultimately need to be transferred.

Images