WO2023035750A1 - 干扰优化方法及装置、存储介质及电子设备 - Google Patents

干扰优化方法及装置、存储介质及电子设备 Download PDF

Info

Publication number
WO2023035750A1
WO2023035750A1 PCT/CN2022/103772 CN2022103772W WO2023035750A1 WO 2023035750 A1 WO2023035750 A1 WO 2023035750A1 CN 2022103772 W CN2022103772 W CN 2022103772W WO 2023035750 A1 WO2023035750 A1 WO 2023035750A1
Authority
WO
WIPO (PCT)
Prior art keywords
station
interference
disturbed
base station
data
Prior art date
Application number
PCT/CN2022/103772
Other languages
English (en)
French (fr)
Inventor
袁晶晶
许森
张乐
Original Assignee
中国电信股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 中国电信股份有限公司 filed Critical 中国电信股份有限公司
Publication of WO2023035750A1 publication Critical patent/WO2023035750A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/08—Testing, supervising or monitoring using real traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/10—Scheduling measurement reports ; Arrangements for measurement reports
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Definitions

  • the present disclosure relates to the field of computer technology, and in particular to an interference optimization method and device, a storage medium, and electronic equipment.
  • TDD Time Division Duplexing, time division duplex
  • the base station side can actively identify the problem, and then the disturbed side can determine the interference source, and then relevant staff can manually configure the optimization strategy.
  • an interference optimization method including:
  • Collect index data call the clustering algorithm constructed by the MDA (Management Data Analytics, management data analysis) function, judge whether there is interference based on the index data, and determine the disturbed station and the disturbed cell in the case of interference;
  • MDA Management Data Analytics, management data analysis
  • the MDA function is called to determine the macro base station, and the disturbing station is determined according to the disturbed cell and the macro base station;
  • the reinforcement learning algorithm constructed by calling the MDA function is based on the disturbed station information of the disturbed station and the disturbing station information of the disturbing station
  • An interference optimization strategy is generated; and the interference optimization strategy is delivered to the interfering station and/or the interfered station, so as to realize interference optimization.
  • the indicator data includes: MR (Measurement Report, measurement report) data reported by the terminal to the base station and base station performance data reported by the base station; wherein, the basic fields of the MR data include: the base station identifier and the base station cell identifier; the MR data
  • the dedicated fields include: RSRP (Reference Signal Receiving Power, reference signal received power); the basic fields of base station performance data include: base station identification and base station cell identification; the exclusive fields of base station performance data include: uplink throughput, uplink PRB (Physical Resource Block , physical resource module) to schedule resources; and, after collecting the index data, further includes: associating and storing the MR data and the base station performance data through the same basic field.
  • calling the clustering algorithm constructed by the MDA function judging whether there is interference based on the index data, and determining the disturbed station and the disturbed cell in the case of interference, includes: calling the MDA function to obtain the index data, extracting The PRB scheduling resources, RSRP and uplink throughput fields in the index data are used to form a three-dimensional array; call the MDA function, use the k-means algorithm to perform cluster analysis on the three-dimensional array, and obtain the clustering result; clustering data with pre-set conditions; if it exists, determine that there is interference; and after determining that there is interference, determine the disturbed station and the disturbed cell from the clustering data that meet the preset conditions.
  • the MDA function is called, and the k-means (k average) algorithm is used to carry out cluster analysis on the three-dimensional arrays to obtain the clustering results, including: taking the number of three-dimensional arrays as the observation value p, and dividing p three-dimensional arrays Divided into k initial sets; among them, k ⁇ p; clustering with the goal of minimizing the sum of squares in the cluster to obtain k result sets as the clustering result.
  • judging whether there is clustering data satisfying preset conditions in the clustering results includes: calculating the average value of PRB scheduling resources, the average value of RSRP and the average value of uplink throughput of each result set in the k result sets ; According to the average value of PRB scheduling resources, the average value of RSRP and the average value of uplink throughput of each two result sets, determine the mean value difference of PRB scheduling resources, the mean value difference of RSRP and the mean value difference of uplink throughput of each two result sets; Compare the PRB scheduling resource mean difference, RSRP mean difference, and uplink throughput mean difference between each two result sets with the PRB scheduling resource threshold, RSRP threshold, and uplink throughput threshold in the preset conditions, and determine whether There are PRB scheduling resource mean difference, RSRP mean difference and uplink throughput mean difference satisfying the inequality relationship in the preset condition; if they exist, they will be compared with the PRB scheduling resource mean difference, RSRP The result set related to the mean difference value and the uplink throughput mean
  • the uplink throughput threshold is greater than 3 times the PRB scheduling resource threshold, and the uplink throughput threshold is greater than 3 times the RSRP threshold.
  • determining the disturbed station and the disturbed cell from the clustering data satisfying the preset condition includes: for each clustering data composed of two result sets satisfying the preset condition, calculating the uplink throughput The result set with the smaller average value is used as the target set; the base station and the cell in the target set are respectively regarded as the disturbed station and the disturbed cell.
  • calling the MDA function to determine the macro base station includes: calling the MDA function to extract the base station latitude and longitude data from the base station parameters; based on the base station latitude and longitude data, determine the base station whose distance from the disturbed station is less than the distance threshold as the macro base station.
  • determining the interfering station according to the disturbed cell and the macro base station includes: calling the MDA function to extract RIM-RS (Remote Interference Management-Reference Signal, remote interference management reference signal) information of the macro base station; wherein, the macro base station
  • the RIM-RS information includes: macro base station cell identifier, macro base station beam identifier, and interference information sent by the interfered cell; based on the RIM-RS information, perform correlation analysis on the disturbed cell, and determine the interfering station from the macro base station; among them,
  • the interfering station information of the interfering station includes: an interfering station identifier, an interfering cell identifier, and an interfering beam identifier.
  • the reinforcement learning algorithm constructed by the MDA function is invoked, and the interference optimization strategy is generated based on the disturbed station information of the disturbed station and the disturbing station information of the disturbing station, including: constructing an objective function and corresponding constraints; , the objective function is the maximization function of the weighted value of the throughput of all users under the disturbed station and the throughput of all users under the disturbed station; using the constructed deep reinforcement learning algorithm, the interference optimization strategy is solved according to the objective function and constraints .
  • the interference optimization strategy is solved according to the objective function and constraints, including: configuring action sets and state sets;
  • the action set includes: the interfering station adjusts the beam SSB direction of the interference, the interfering station reduces the transmission power by 3dB, the interfering station reduces the transmission power by 6dB, closes the time slot resources for the interfering station to generate interference, and closes the interfered station.
  • the time slot resource where the station interferes; the state set includes: the performance data reported by the disturbed station and the PRB scheduling resource scheduled in the MR data, RSRP and uplink throughput.
  • the interference optimization strategy after delivering the interference optimization strategy to the interfering station and/or the disturbed station, it further includes: collecting optimized index data of the disturbed station and the disturbed cell, calling the MDA function, and based on the optimized index The data judges whether there is interference to achieve optimal feedback.
  • an interference optimization device including:
  • the data acquisition module is used to collect index data; the judgment victim module is used to call the clustering algorithm constructed by the MDA function, and judges whether there is interference based on the index data; the module of determining the disturbed station is used for when there is interference Determine the disturbed station and the disturbed cell; determine the disturbing station module, used to call the MDA function to determine the macro base station in the case of interference, and determine the disturbed station according to the disturbed cell and the macro base station; generate a strategy module for Invoke the reinforcement learning algorithm constructed by the MDA function to generate an interference optimization strategy based on the disturbed station information of the disturbed station and the disturbing station information of the disturbing station; the delivery module is used to send the interference optimization strategy to the disturbing station and/or or victim station for interference optimization.
  • the indicator data includes: MR data reported by the terminal to the base station and base station performance data reported by the base station; wherein, the basic fields of the MR data include: the base station identifier and the base station cell identifier; the dedicated fields of the MR data include: RSRP; The basic fields of performance data include: base station identification and base station cell identification; the dedicated fields of base station performance data include: uplink throughput, uplink PRB scheduling resources; and, after collecting index data, also include: MR data and base station performance data through The same underlying fields are stored associatively.
  • the disturbed module calls the clustering algorithm constructed by the MDA function, judges whether there is interference based on the index data, and determines the disturbed station and the disturbed cell in the case of interference, including: calling the MDA function to obtain Index data, extract the PRB scheduling resources, RSRP and uplink throughput fields in the index data to form a three-dimensional array; call the MDA function, use the k-means algorithm to perform cluster analysis on the three-dimensional array, and obtain the clustering result; judge the clustering result Whether there is clustering data satisfying the preset condition; if so, determining that there is interference; and determining the disturbed station and the disturbed cell from the clustering data satisfying the preset condition after determining that there is interference.
  • the disturbed module calls the MDA function, uses the k-means algorithm to perform cluster analysis on the three-dimensional arrays, and obtains the clustering results, including: taking the number of three-dimensional arrays as the observation value p, and dividing p three-dimensional arrays Divided into k initial sets; among them, k ⁇ p; clustering with the goal of minimizing the sum of squares in the cluster to obtain k result sets as the clustering result.
  • the judging disturbed module judges whether there is clustering data that satisfies preset conditions in the clustering results, including: calculating the average value of PRB scheduling resources, the average value of RSRP, and the uplink Throughput average: According to the average value of PRB scheduling resources, average value of RSRP, and average value of uplink throughput of each two result sets, determine the difference between the average value of PRB scheduling resources, the difference between the average value of RSRP, and the uplink throughput of each two result sets Mean difference: the PRB scheduling resource mean difference, RSRP mean difference and uplink throughput mean difference of each two result sets are compared with the PRB scheduling resource threshold, RSRP threshold and uplink throughput threshold in the preset conditions respectively Compare and judge whether there is a PRB scheduling resource mean difference, RSRP mean difference, and uplink throughput mean difference that satisfy the inequality relationship in the preset condition; A result set related to the difference, the RSRP mean difference, and the uplink throughput mean difference is determined to be the clustering data satisfying the
  • the uplink throughput threshold is greater than 3 times the PRB scheduling resource threshold, and the uplink throughput threshold is greater than 3 times the RSRP threshold.
  • the module for determining the disturbed station determines the disturbed station and the disturbed cell from the cluster data satisfying the preset condition, including: for each cluster data composed of two result sets satisfying the preset condition, The result set in which the average uplink throughput is smaller is taken as the target set; the base station and the cell in the target set are respectively taken as the disturbed station and the disturbed cell.
  • the determination of the interfering station module calling the MDA function to determine the macro base station includes: calling the MDA function to extract the base station latitude and longitude data from the base station parameters;
  • the base station serves as a macro base station.
  • the module for determining the interfering station determines the interfering station according to the disturbed cell and the macro base station, including: calling the MDA function to extract the remote interference management reference signal RIM-RS information of the macro base station; wherein, the RIM-RS of the macro base station
  • the information includes: the macro base station cell ID, the macro base station beam ID, and the interference information sent by the interfered cell; based on RIM-RS information, correlation analysis is performed on the interfered cell, and the interfering station is determined from the macro base station; among them, the interfering station's
  • the information of the interfering station includes: the identification of the interfering station, the identification of the interfering cell, and the identification of the interfering beam.
  • the generation strategy module invokes the reinforcement learning algorithm constructed by the MDA function, and generates an interference optimization strategy based on the disturbed station information of the disturbed station and the disturbing station information of the disturbing station, including: constructing an objective function and corresponding constraints conditions; where, the objective function is the maximization function of the weighted value of the throughput of all users under the disturbed station and the throughput of all users under the disturbed station; using the constructed deep reinforcement learning algorithm, according to the objective function and constraints, the interfere with optimization strategies.
  • the generation strategy module uses the constructed deep reinforcement learning algorithm to solve the interference optimization strategy according to the objective function and constraints, including: configuring action sets and state sets;
  • the action set includes: the interfering station adjusts the beam SSB direction of the interference, the interfering station reduces the transmission power by 3dB, the interfering station reduces the transmission power by 6dB, closes the time slot resources for the interfering station to generate interference, and closes the interfered station.
  • the time slot resource where the station interferes; the state set includes: the performance data reported by the disturbed station and the PRB scheduling resource scheduled in the MR data, RSRP and uplink throughput.
  • the sending module is further configured to: after sending the interference optimization strategy to the disturbing station and/or the disturbed station, further include: collecting optimized index data of the disturbed station and the disturbed cell, calling The MDA function judges whether there is interference based on the optimized index data to achieve optimization feedback.
  • a non-volatile computer-readable storage medium on which a computer program is stored, and the above-mentioned interference optimization method is implemented when the computer program is executed by a processor.
  • an electronic device including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the executable instructions to Perform the interference optimization method described above.
  • a computer program including: instructions, which when executed by a processor cause the processor to execute the interference optimization method of any one of the above embodiments.
  • a computer program product including instructions, which when executed by a processor cause the processor to execute the interference optimization method of any one of the above embodiments.
  • FIG. 1 shows a schematic diagram of an exemplary system architecture to which the interference optimization method of some embodiments of the present disclosure can be applied;
  • FIG. 2 shows a flowchart of an interference optimization method in some embodiments of the present disclosure
  • FIG. 3 shows a flowchart of a method for judging whether there is interference in some embodiments of the present disclosure
  • FIG. 4 shows a flowchart of a method for determining a disturbed station in some embodiments of the present disclosure
  • Fig. 5 shows a flowchart of a method for determining an interfering station in the presence of interference according to some embodiments of the present disclosure
  • FIG. 6 shows a flowchart of a method for generating an interference optimization strategy in some embodiments of the present disclosure
  • Fig. 7 shows a schematic diagram of an apparatus for implementing interference optimization according to some embodiments of the present disclosure
  • FIG. 8 shows a block diagram of an interference optimization device of some embodiments of the present disclosure.
  • Fig. 9 shows a structural block diagram of an interference optimization computer device in some embodiments of the present disclosure.
  • Example embodiments will now be described more fully with reference to the accompanying drawings.
  • Example embodiments may, however, be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete and will fully convey the concept of example embodiments to those skilled in the art.
  • the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
  • first and second are used for descriptive purposes only, and cannot be interpreted as indicating or implying relative importance or implicitly specifying the quantity of indicated technical features. Thus, a feature defined as “first” and “second” may explicitly or implicitly include one or more of these features.
  • “plurality” means at least two, such as two, three, etc., unless otherwise specifically defined.
  • the 3GPP TR28.809 agreement determines the role and function of MDAS (Management Data Analytics service) in the network management system in the management cycle.
  • MDAS Management Data Analytics service
  • the data analysis management service can identify network performance degradation problems based on machine learning models, and Accurately identify the root cause of the problem, optimize the strategy to solve the problem, determine the interactive feedback mechanism between the user and the provider, and the data input information and output attributes of relevant scenarios.
  • problems in the current standard and implementation 1) There is a lack of full control management and specific analysis methods on the network management side.
  • the solution to the cross-interference problem of the TDD frequency band flexible duplex networking configuration scheme is to actively identify the interference problem on the base station side, send a signal to the disturbed station to determine the interference source, and then manually obtain the interference optimization strategy configured by the network management.
  • the calculation complexity on the base station side is increased, and the method lacks flexibility.
  • there will be information loss in the information transmission between the disturbed station and the disturbing station which will reduce the timeliness of solving the interference problem, and intermittently affect the user's data quality; 2)
  • the network management side lacks specific data for this scenario Processing mechanism: To realize the full control mechanism of the network management, it is necessary to use the MDA (Management Data Analytics, management data analysis) function to identify the interference problem and obtain the interference optimization strategy.
  • MDA Management Data Analytics, management data analysis
  • the centralized data acquisition unit on the RAN (Radio Access Network, wireless access network) side carries out corresponding index data collection and data analysis direction, and the demand for output attributes cannot be determined, which affects the realization of the intelligent solution for interference optimization in the above scenarios.
  • the computational complexity of the base station side is relatively high, and information transmission between the victim station and the disturber station often has information loss, which reduces the timeliness of solving interference problems and affects user data. quality.
  • exemplary embodiments of the present disclosure provide an interference optimization method for at least solving one or all of the above technical problems.
  • the purpose of the present disclosure is to provide an interference optimization method, device, electronic equipment and storage medium, so as to provide an interference optimization method executed by the network management system.
  • the data analysis and management service MDAS mechanism is introduced to confirm the interference. Phenomena, determine the disturbed station, determine the disturbing station, generate an interference optimization strategy and send it to the disturbed station and/or the disturbing station, and then realize the fully closed-loop process of interference optimization led by the network management side.
  • FIG. 1 shows a schematic diagram of an exemplary system architecture to which the interference optimization method of some embodiments of the present disclosure can be applied; as shown in FIG. 1 :
  • the system architecture may include a server 101 , a network 102 and a client 103 .
  • the network 102 is used as a medium for providing a communication link between the client 103 and the server 101 .
  • Network 102 may include various connection types, such as wires, wireless communication links, or fiber optic cables, among others.
  • the server 101 can provide services for network management control, and can be a server that provides various services, such as providing index data reported by the receiving terminal (client 103), judging whether there is interference phenomenon according to the index data, and determining the interference phenomenon when there is interference phenomenon.
  • the backend management server for functional services such as station and interference station, generation of interference optimization strategy and delivery. That is: the background management server installed in the network management system can confirm whether there is interference phenomenon according to the index data reported by the client 103, determine the disturbed station and the disturbing station, generate and issue an interference optimization strategy, so as to realize the interference optimization dominated by the network management side
  • a fully closed-loop process improves the data quality of the client 103.
  • the client 103 can be a mobile terminal such as a mobile phone, a game host, a tablet computer, an e-book reader, smart glasses, a smart home device, an AR (Augmented Reality, augmented reality) device, a VR (Virtual Reality, virtual reality) device, or,
  • the client 103 may also be a personal computer, such as a laptop computer, a desktop computer, and the like.
  • the server 101 can call the clustering algorithm constructed by the management data analysis MDA function to judge whether there is interference according to the indicator data, and if there is interference, Determine the disturbed station and the disturbed cell; you can also call the MDA function to determine the macro base station, so as to determine the disturbing station according to the disturbed cell and the macro base station; then call the reinforcement learning algorithm constructed by the MDA function, based on the disturbed station.
  • the information and the information of the interfering station of the interfering station generate an interference optimization strategy, so as to deliver the interference optimization strategy to the interfering station and/or the interfered station, thereby realizing interference optimization.
  • the server 101 may also collect optimized index data reported by the disturbed station and the client 103 under the disturbed cell after delivering the interference optimization strategy, so as to realize optimization feedback.
  • the numbers of clients, networks and servers in FIG. 1 are only illustrative, and the server 101 can be a physical server, a server cluster composed of multiple servers, or a cloud server, depending on actual needs , which can have any number of clients, networks, and servers.
  • Fig. 2 shows a flowchart of an interference optimization method in some embodiments of the present disclosure.
  • the method provided by the embodiment of the present disclosure may be executed by the server or the client as shown in FIG. 1 , but the present disclosure is not limited thereto.
  • server cluster 101 is used as the execution subject for illustration.
  • the interference optimization method provided by the embodiment of the present disclosure can be executed by the network management system, including the following steps:
  • Step S201 collect index data.
  • the network management system can collect and monitor the MR (Measurement Report, measurement report) data reported by the terminal to the base station and the performance data reported by the base station within a certain period of time.
  • the index data can include: the measurement report MR data reported by the terminal to the base station and The base station performance data reported by the base station; wherein, the basic fields of MR data include: base station identification and base station cell identification; the dedicated fields of MR data include: RSRP (Reference Signal Received Power, reference signal received power); the basic fields of base station performance data include : base station identification and base station cell identification; the dedicated fields of base station performance data include: uplink throughput, uplink physical resource block PRB (Physical Resource Block, physical resource block) scheduling resources; and, after collecting index data, MR data can also be It is stored in association with base station performance data through the same basic field.
  • PRB Physical Resource Block
  • Step S203 calling the clustering algorithm constructed by the management data analysis MDA function, judging whether there is interference based on the index data, and determining the disturbed station and the disturbed cell if there is interference.
  • Step S205 in the case of interference, call the MDA function to determine the macro base station, and determine the interfering station according to the disturbed cell and the macro base station.
  • the reinforcement learning algorithm constructed by the MDA function is invoked to generate an interference optimization strategy based on the information of the victim station of the victim station and the information of the aggressor station of the aggressor station.
  • Step S209 delivering the interference optimization strategy to the interfering station and/or the interfered station, so as to realize interference optimization.
  • This application proposes an interference optimization method based on network management data analysis and processing executed by the network management system.
  • the main idea is: for the TDD flexible duplex networking scenario deployed by 5G large uplink users, the network management system collects the MR data reported by the terminal and the base station in real time Reported base station performance and other index data, call the clustering algorithm constructed by the MDA function to analyze PRB, RSRP and uplink throughput data to actively identify the existence of interference and find out the disturbed station; and when interference is found, it can analyze the The RIM-RS signal information of the macro base stations around the disturbing station is used to locate and discover the disturbing station; then the reinforcement learning algorithm constructed by MDA is used to analyze and generate the interference optimization strategy, so as to send the configuration information of the interference optimization strategy to the corresponding disturbing station or the disturbed station (to) complete the interference optimization process.
  • the network management system utilizes the management data analysis MDA function to realize the fully controlled closed-loop process of identification of interference problems, identification of disturbed stations and interference sources, and generation of interference optimization strategies.
  • calling the clustering algorithm constructed by the management data analysis MDA function judging whether there is interference based on the index data, and determining the disturbed station and the disturbed cell in the case of interference, includes: calling the MDA function to obtain the index Data, extract the PRB scheduling resources, RSRP and uplink throughput fields in the index data to form a three-dimensional array; call the MDA function, use the k-means algorithm to perform cluster analysis on the three-dimensional array, and obtain the clustering result; determine whether the clustering result is There is clustering data satisfying the preset condition; if there is, determining that there is interference; and after determining that there is interference, determining the disturbed station and the disturbed cell from the clustering data satisfying the preset condition.
  • the interference optimization method provided by the embodiments of the present disclosure can confirm the interference phenomenon, determine the disturbed station, determine the disturbing station, generate an interference optimization strategy and send it to The victim station and/or the disturber station, so as to realize the fully closed-loop process of interference optimization led by the network management side.
  • Fig. 3 shows a flowchart of a method for judging whether there is interference in some embodiments of the present disclosure, as shown in Fig. 3 , including:
  • Step 301 calling the MDA function to obtain index data, extracting PRB scheduling resources, RSRP and uplink throughput fields in the index data to form a three-dimensional array;
  • Step 303 calling the MDA function, using the k-means algorithm to perform cluster analysis on the three-dimensional array to obtain the clustering result;
  • Step 305 judging whether there is clustering data satisfying the preset condition in the clustering results; if so, judging in step 307 that there is interference; if not, judging in step 309 that there is no interference.
  • the network management system can call the MDA function, use the k-means algorithm to perform cluster analysis on the three-dimensional array, and obtain the clustering result, including: taking the number of three-dimensional arrays as the observation value p, and dividing p The three-dimensional array is divided into k initial sets; among them, k ⁇ p; clustering is performed with the goal of minimizing the sum of squares within the cluster to obtain k result sets as the clustering result.
  • the MDA function of the network management system can be used to analyze the data to identify whether interference exists.
  • ⁇ i is the cluster center point of each group.
  • a clustering group identifier (such as: 1, 2, 3...k) can be added behind each group of data to output as the final clustering result.
  • judging whether there is clustering data satisfying preset conditions in the clustering results includes: calculating the average value of PRB scheduling resources, the average value of RSRP and the average value of uplink throughput of each result set in the k result sets ; According to the average value of PRB scheduling resources, the average value of RSRP and the average value of uplink throughput of each two result sets, determine the mean value difference of PRB scheduling resources, the mean value difference of RSRP and the mean value difference of uplink throughput of each two result sets; Compare the PRB scheduling resource mean difference, RSRP mean difference, and uplink throughput mean difference between each two result sets with the PRB scheduling resource threshold, RSRP threshold, and uplink throughput threshold in the preset conditions, and determine whether There are PRB scheduling resource mean difference, RSRP mean difference and uplink throughput mean difference satisfying the inequality relationship in the preset condition; if they exist, they will be compared with the PRB scheduling resource mean difference, RSRP The result set related to the mean difference value and the uplink throughput mean
  • Avg prb_i and Avg prb_j respectively represent the average value of PRB scheduling resources of the two result sets;
  • Avg rsrp_i and Avg rsrp_j represent the average RSRP values of the two result sets respectively;
  • Avg thp_i and Avg thp_j respectively represent the average uplink throughput of the two result sets;
  • A represents the PRB scheduling resource threshold, B represents the RSRP threshold, and C represents the uplink throughput threshold.
  • the values of A and B can be determined by referring to actual applications. OK, the relationship between A, B and C should hold: C is more than 3 times that of A, and C is more than 3 times that of B.
  • determining the disturbed station and the disturbed cell from the clustering data satisfying the preset condition includes: for each clustering data composed of two result sets satisfying the preset condition, calculating the uplink throughput The result set with the smaller average value is used as the target set; the base station and the cell in the target set are respectively regarded as the disturbed station and the disturbed cell.
  • Fig. 4 shows a flowchart of a method for determining a disturbed station in some embodiments of the present disclosure, as shown in Fig. 4 , including:
  • Step 401 calculating the average value of PRB scheduling resources, the average value of RSRP and the average value of uplink throughput in each of the k result sets;
  • Step 403 according to the average value of PRB scheduling resources, average value of RSRP, and average value of uplink throughput of each two result sets, determine the difference between the average value of PRB scheduling resources, the average value of RSRP, and the average value of uplink throughput of each two result sets value;
  • Step 405 Compare the PRB scheduling resource mean difference, RSRP mean difference, and uplink throughput mean difference between each two result sets with the PRB scheduling resource threshold, RSRP threshold, and uplink throughput threshold in preset conditions, respectively ;
  • Step 407 when there is a PRB scheduling resource mean difference, RSRP mean difference, and uplink throughput mean difference satisfying the inequality relationship in the preset condition, compare it with the PRB scheduling resource mean difference, The result set related to the mean difference of RSRP and the mean difference of uplink throughput is determined as the clustering data satisfying the preset condition;
  • Step 409 for each clustering data composed of two result sets satisfying the preset condition, use the result set whose uplink throughput average value is smaller as the target set;
  • Step 411 taking the base station and the cell in the target set as the disturbed station and the disturbed cell respectively.
  • calling the MDA function to determine the macro base station includes: calling the MDA function to extract the base station latitude and longitude data from the base station parameters; based on the base station latitude and longitude data, determine the base station whose distance from the disturbed station is less than the distance threshold as the macro base station.
  • the network management system can extract the latitude and longitude data in the base station parameters, and then obtain the latitude and longitude data of the disturbed station and the latitude and longitude data of the first macro base station within the surrounding range of the disturbed station, and then pass the distance between the disturbed station and the first macro base station , find the first macro base station whose distance S to the victim station is within the threshold range D, as the macro base station in the above embodiment.
  • the calculation method of calculating the distance S between them is as follows:
  • determining the interfering station according to the disturbed cell and the macro base station includes: calling the MDA function to extract the remote interference management reference signal RIM-RS information of the macro base station; wherein, the RIM-RS information of the macro base station includes: the macro base station The cell ID, the beam ID of the macro base station, and the interference information sent by the interfered cell; based on the RIM-RS information, correlation analysis is performed on the interfered cell, and the interfering station is determined from the macro base station; the interfering station information of the interfering station includes : The ID of the interfering station, the ID of the interfering cell, and the ID of the interfering beam.
  • Fig. 5 shows a flowchart of a method for determining an interfering station in the presence of interference according to some embodiments of the present disclosure, as shown in Fig. 5 , including:
  • Step 501 calling the MDA function to extract the base station latitude and longitude data from the base station industrial parameters; in some practical applications, the base station latitude and longitude data can also be extracted from other data tables that record the latitude and longitude information;
  • Step 503 determining the latitude and longitude data of the disturbed station according to the latitude and longitude data of the base station, and the latitude and longitude data of the first macro base station around the disturbed station;
  • Step 505 calculating the distance S between the disturbed station and the first macro base station, and finding the first macro base station corresponding to S that is smaller than the distance threshold D as the macro base station;
  • Step 507 calling the MDA function to extract the RIM-RS information of the remote interference management reference signal of the macro base station;
  • the RIM-RS information of the macro base station includes: the cell identifier of the macro base station, the beam identifier of the macro base station, and the interference information sent by the interfered cell;
  • Step 509 Carry out correlation analysis on the disturbed cell based on the RIM-RS information, and determine the disturbing station from the macro base station; wherein, the disturbing station information of the disturbing station includes: the identity of the disturbing station, the identity of the disturbing cell, and the identity of the disturbing station. Beam ID.
  • the reinforcement learning algorithm constructed by the MDA function is invoked, and the interference optimization strategy is generated based on the disturbed station information of the disturbed station and the disturbing station information of the disturbing station, including: constructing an objective function and corresponding constraints; , the objective function is the maximization function of the weighted value of the throughput of all users under the disturbed station and the throughput of all users under the disturbed station; using the constructed deep reinforcement learning algorithm, the interference optimization strategy is solved according to the objective function and constraints .
  • the network management MDA function in the network management system of MDAS is introduced, and the real-time online generation of interference optimization strategies can be realized through the establishment of a good algorithm model.
  • the meaning of the objective function can be set to maximize the weighted value of the throughput of all users under the disturbed station and the throughput of all users under the disturbing station per unit time;
  • the objective function can be set as follows:
  • Thp(t) represents the throughput weighted value of all users under the disturbed station and the throughput of all users under the disturbing station per unit time; N represents the total number of all users under the disturbed station, and M represents the throughput of all users under the disturbing station. The total number of all users.
  • the constraint effect of the constraint condition can be set as follows: to ensure the QoS (Quality of Service, service quality) of the user at the disturbed station as much as possible, and reduce the performance impact on the user at the disturbing station at the same time; the constraint condition can be set as follows:
  • any two result sets S i and S j can be determined first, and Avg prb_i and Avg prb_j in the constraints represent the average value of PRB scheduling resources of these two result sets respectively; Avg rsrp_i and Avg rsrp_j represent the average RSRP values of the two result sets respectively; Avg thp_i and Avg thp_j represent the average uplink throughput of the two result sets respectively; where A represents the PRB scheduling resource threshold, B represents the RSRP threshold, and C represents The uplink throughput threshold, the values of A and B can be determined with reference to the actual application, the relationship between A, B and C should keep C being more than 3 times of A, and C being more than 3 times of B. Then the setting of the first condition in the constraint conditions can make each result set prb, throughput, and RSRP in the clustering results after the clustering in the preceding steps are all within the corresponding threshold;
  • the setting of the second condition in the constraints can make the total throughput of all users under the disturbed station not be lower than ⁇ 0 ,
  • the setting of the second condition in the constraint conditions can make the probability that the total throughput of all users under the disturbing station is lower than q is less than ⁇ , where the value of q can be determined according to the user load m under the disturbing station, and the relationship into a negative correlation:
  • the interference optimization strategy is solved according to the objective function and constraints, including: configuring action sets and state sets;
  • the three-layer network structure can include: the input layer of neuron+rule function is set, the intermediate layer of neuron+relu function is set and the output layer of two functions of neuron+sigmoid+softmax is set.
  • the action set includes: the interfering station adjusts the beam SSB direction of the interference, the interfering station reduces the transmission power by 3dB, the interfering station reduces the transmission power by 6dB, closes the time slot resources for the interfering station to generate interference, and closes the interfered station.
  • the time slot resource where the station generates interference; the data in the action set can be considered as the transfer action of five attributes.
  • the state set includes: the performance data reported by the disturbed station and the PRB scheduling resource, RSRP and uplink throughput scheduled in the MR data.
  • the policy generation can be performed based on the output results of the previous steps, which can include constructing the Markov tuple user state and actions from the current moment to the next moment, and constructing a system reward function from the state and actions, Construct a deep reinforcement learning model to determine the optimal interference optimization strategy.
  • Fig. 6 shows a flowchart of a method for generating an interference optimization strategy in some embodiments of the present disclosure, as shown in Fig. 6 , including:
  • Step S601 constructing an objective function and corresponding constraint conditions
  • Step S603 configuring action sets and state sets, and constructing a policy network and an action evaluation network
  • Step S605 using the policy network to determine a policy based on the action set and the state set based on the policy policy;
  • Step S607 use the action evaluation network to solve the strategy convergence, and obtain the interference optimization strategy.
  • the configuration information of the interference optimization strategy can be obtained through the reinforcement learning algorithm constructed by the MDA function, and the configuration information can indicate the adjustment strategy related to the five transfer actions in the action set; the interference optimization strategy is sent to the interference station and/or the victim station, the interference station and/or the victim station can make specific adjustments based on the configuration information in the policy, that is, the specific output attributes can be obtained through the reinforcement learning algorithm constructed by the MDA function, which is used to use The interfering station and/or victim station perform to resolve the interference problem.
  • the interference optimization strategy after delivering the interference optimization strategy to the interfering station and/or the disturbed station, it further includes: collecting optimized index data of the disturbed station and the disturbed cell, calling the MDA function, and based on the optimized index The data judges whether there is interference to achieve optimal feedback.
  • the network management device can be used to send the interference configuration information generated by the MDA function to the interference station and/or the interference station to take effect.
  • the NMS collects the optimized disturbed-side data in real time, determines whether the disturbed-side interference information is resolved, and iteratively performs MDA processing and analysis in real time.
  • the method in this application has the function of verification feedback, MDA updates the process of model parameters and strategies online in real time, iteratively optimizes strategies, and then produces better optimization results.
  • Fig. 7 shows a schematic diagram of an apparatus for implementing interference optimization in some embodiments of the present disclosure, as shown in Fig. 7 , including:
  • the introduced MDA function can be used to implement functions such as data collection and classification, identification and analysis of interference problems, identification and analysis of interference stations, and generation and analysis of interference optimization strategies. It will take effect until the interference station and/or the interference station takes effect, and solve the interference problem in real time. In some practical applications, after the interfering station and/or the interfering station implements the interference optimization strategy, the network management device side can also collect the optimized data of the disturbed side, conduct analysis to give feedback on the optimization effect in time, and update the learning model in time to adjust the optimization strategy. scheme to achieve better optimization results.
  • this application has an interactive mechanism for data collection and configuration information distribution, which can update the establishment of the model in the MDA function in real time, optimize the output, and the interaction process between devices is integrated with the existing network management structure, which is convenient for deployment and implementation .
  • the network management side can collect the input for interference optimization in this scenario
  • the data can actively carry out model construction, and can obtain output attributes for optimization through the model; among them, the RAN centralized data collection unit on the network management side can realize index data collection, and the introduced MDA function can realize model construction and data analysis, and then can From the perspective of full control of the network management, the full-control closed-loop process of active identification of interference problems, determination of interference sources, and generation of interference optimization strategies can be realized, thereby ensuring the real-time identification of interference problems in this scenario and the efficiency of generating interference strategies.
  • the network management system proposed in this application adds the MDA function of this scenario, which can realize data collection and input, analysis mechanism, and output attributes from the network management side, and operators can implement data collection and effective data collection in a targeted manner.
  • the purposeful development of model algorithms provides basic support for operators to deploy network optimization and upgrades for industries with high uplink requirements for vertical industries.
  • the interference optimization method provided by this application can also reduce the computational complexity of the base station, perform data processing and analysis from the perspective of the network management, and automatically deliver the interference optimization strategy configuration information to take effect, reducing the burden of manual configuration of the operation and maintenance personnel and improving the operation and maintenance. efficiency.
  • FIG. 8 shows a block diagram of an interference optimization device 800 in some embodiments of the present disclosure; as shown in FIG. 8 , it includes:
  • Data collection module 801 used to collect index data
  • Judging the disturbed module 802 used to call the clustering algorithm of the management data analysis MDA function structure, and judge whether there is interference based on the index data;
  • the delivery module 806 is configured to deliver the interference optimization policy to the interfering station and/or the interfered station, so as to realize interference optimization.
  • Fig. 9 shows a structural block diagram of an interference optimization computer device in some embodiments of the present disclosure. It should be noted that the electronic device shown in the figure is only an example, and should not limit the functions and application scope of the embodiments of the present disclosure.
  • FIG. 9 An electronic device 900 according to this embodiment of the present disclosure is described below with reference to FIG. 9 .
  • the electronic device 900 shown in FIG. 9 is only an example, and should not limit the functions and scope of use of the embodiments of the present disclosure.
  • electronic device 900 takes the form of a general-purpose computing device.
  • Components of the electronic device 900 may include but not limited to: at least one processing unit 910 , at least one storage unit 920 , and a bus 930 connecting different system components (including the storage unit 920 and the processing unit 910 ).
  • the storage unit stores program codes, and the program codes can be executed by the processing unit 910, so that the processing unit 910 executes various exemplary methods according to the present disclosure described in the "Exemplary Methods" section of this specification.
  • the processing unit 910 may execute step S201 as shown in FIG. 2 to collect index data; step S203 to call the clustering algorithm constructed by the management data analysis MDA function, judge whether there is interference based on the index data, and determine whether there is interference when there is interference. In the case of , the disturbed station and the disturbed cell are determined.
  • Step S205 in the case of interference, call the MDA function to determine the macro base station, and determine the interfering station according to the disturbed cell and the macro base station.
  • step S207 the reinforcement learning algorithm constructed by the MDA function is invoked to generate an interference optimization strategy based on the information of the victim station of the victim station and the information of the aggressor station of the aggressor station.
  • step S209 the interference optimization strategy is issued to the interfering station and/or the interfered station, so as to realize interference optimization.
  • the storage unit 920 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 9201 and/or a cache storage unit 9202 , and may further include a read-only storage unit (ROM) 9203 .
  • RAM random access storage unit
  • ROM read-only storage unit
  • the storage unit 920 may also include a program/utility 9204 having a set (at least one) of program modules 9205, such program modules 9205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, Implementations of networked environments may be included in each or some combination of these examples.
  • Bus 930 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local area using any of a variety of bus structures. bus.
  • the electronic device 900 may also communicate with one or more external device interference optimization apparatuses 800 (such as keyboards, pointing devices, Bluetooth devices, etc.), and may also communicate with one or more devices that enable the user to interact with the electronic device 900, and and/or communicate with any device (eg, router, modem, etc.) that enables the electronic device 900 to communicate with one or more other computing devices. Such communication may occur through input/output (I/O) interface 950 .
  • the electronic device 900 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN) and/or a public network such as the Internet) through the network adapter 960 .
  • networks such as a local area network (LAN), a wide area network (WAN) and/or a public network such as the Internet
  • the network adapter 960 communicates with other modules of the electronic device 900 through the bus 930 .
  • other hardware and/or software modules may be used in conjunction with electronic device 900, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives And data backup storage system, etc.
  • the example implementations described here can be implemented by software, or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of software products, and the software products can be stored in a non-volatile storage medium (which can be CD-ROM, U disk, mobile hard disk, etc.) or on the network , including several instructions to make a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) execute the method according to the embodiments of the present disclosure.
  • a computing device which may be a personal computer, a server, a terminal device, or a network device, etc.
  • a computer-readable storage medium on which a program product capable of implementing the above-mentioned method in this specification is stored.
  • various aspects of the present disclosure may also be implemented in the form of a program product, which includes program code, and when the program product is run on a terminal device, the program code is used to make the The terminal device executes the steps according to various exemplary embodiments of the present disclosure described in the "Exemplary Method" section above in this specification.
  • a computer program including: instructions, which, when executed by a processor, cause the processor to execute the interference optimization method of any one of the above embodiments.
  • a computer program product including instructions, which, when executed by a processor, cause the processor to execute the interference optimization method of any one of the above embodiments.
  • a program product for implementing the above method according to the embodiment of the present disclosure, it may adopt a portable compact disk read only memory (CD-ROM) and include program codes, and may run on a terminal device such as a personal computer.
  • CD-ROM compact disk read only memory
  • the program product of the present disclosure is not limited thereto.
  • a readable storage medium may be any tangible medium containing or storing a program, and the program may be used by or in combination with an instruction execution system, apparatus or device.
  • the program product may reside on any combination of one or more readable media.
  • the readable medium may be a readable signal medium or a readable storage medium.
  • the readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: electrical connection with one or more conductors, portable disk, hard disk, random access memory (RAM), read only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
  • a computer readable signal medium may include a data signal carrying readable program code in baseband or as part of a carrier wave. Such propagated data signals may take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.
  • a readable signal medium may also be any readable medium other than a readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.
  • Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
  • Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages—such as Java, C++, etc., as well as conventional procedural Programming language - such as "C" or a similar programming language.
  • the program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server to execute.
  • the remote computing device may be connected to the user computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., using an Internet service provider). business to connect via the Internet).
  • LAN local area network
  • WAN wide area network
  • Internet service provider e.g., a wide area network
  • steps of the methods of the present disclosure are depicted in the drawings in a particular order, there is no requirement or implication that the steps must be performed in that particular order, or that all illustrated steps must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step for execution, and/or one step may be decomposed into multiple steps for execution, etc.
  • the technical solutions according to the embodiments of the present disclosure can be embodied in the form of software products, and the software products can be stored in a non-volatile storage medium (which can be CD-ROM, U disk, mobile hard disk, etc.) or on the network , including several instructions to make a computing device (which may be a personal computer, a server, a mobile terminal, or a network device, etc.) execute the method according to the embodiments of the present disclosure.
  • a non-volatile storage medium which can be CD-ROM, U disk, mobile hard disk, etc.
  • a computing device which may be a personal computer, a server, a mobile terminal, or a network device, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本公开提供了干扰优化方法、装置、电子设备及存储介质，涉及计算机技术领域。该方法包括：采集指标数据；调用网管系统的管理数据分析MDA功能构造的聚类算法，基于指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区；在存在干扰的情况下，调用MDA功能确定宏基站，根据受扰小区和宏基站确定施扰站；调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略；将干扰优化策略下发至施扰站或受扰站中的至少一个，以实现干扰优化。

Description

干扰优化方法及装置、存储介质及电子设备
相关申请的交叉引用
本申请是以CN申请号为202111050926.8，申请日为2021年9月8日的申请为基础，并主张其优先权，该CN申请的公开内容在此作为整体引入本申请中。
技术领域
本公开涉及计算机技术领域，尤其涉及一种干扰优化方法及装置、存储介质及电子设备。
背景技术
在5G技术的发展中，TDD(Time Division Duplexing，时分双工)频段灵活双工组网配置方案的应用可以满足5G行业用户对大上行能力的要求，但在该方案的应用下常常发生交叉干扰问题。
相关技术中可以由基站侧主动识别问题，接着由受扰侧确定干扰源，再由相关工作人员手动配置优化策略。
需要说明的是，在上述背景技术部分公开的信息仅用于加强对本公开的背景的理解，因此可以包括不构成对本领域普通技术人员已知的现有技术的信息。
发明内容
本公开的其他特性和优点将通过下面的详细描述变得显然，或部分地通过本公开的实践而习得。
根据本公开的一个方面，提供一种干扰优化方法，包括：
采集指标数据；调用MDA(Management Data Analytics，管理数据分析)功能构造的聚类算法，基于指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区；在存在干扰的情况下，调用MDA功能确定宏基站，根据受扰小区和宏基站确定施扰站；调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略；将干扰优化策略下发至施扰站和/或受扰站，以实现干扰优化。
在一些实施例中，指标数据包括：终端上报基站的MR(Measurement Report， 测量报告)数据和基站上报的基站性能数据；其中，MR数据的基础字段包括：基站标识和基站小区标识；MR数据的专属字段包括：RSRP(Reference Signal Receiving Power，参考信号接收功率)；基站性能数据的基础字段包括：基站标识和基站小区标识；基站性能数据的专属字段包括：上行吞吐量、上行PRB(Physical Resource Block，物理资源模块)调度资源；以及，在采集指标数据后，还包括：将MR数据和基站性能数据通过相同的基础字段进行关联存储。
在一些实施例中，调用MDA功能构造的聚类算法，基于指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区，包括：调用MDA功能获取指标数据，抽取指标数据中的PRB调度资源、RSRP和上行吞吐量字段以组成三维数组；调用MDA功能，使用k-means算法对三维数组进行聚类分析，得到聚类结果；判断聚类结果中是否存在满足预设条件的聚类数据；若存在，则确定存在干扰；以及在确定存在干扰后，从满足预设条件的聚类数据中确定出受扰站和受扰小区。
在一些实施例中，调用MDA功能，使用k-means(k平均)算法对三维数组进行聚类分析，得到聚类结果，包括：以三维数组的个数作为观察值p，将p个三维数组分为k个初始集合；其中，k≤p；以最小化集群内平方和为目标进行聚类得到k个结果集合，以作为聚类结果。
在一些实施例中，判断聚类结果中是否存在满足预设条件的聚类数据，包括：计算k个结果集合中每个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值；根据每两个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值，确定每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值；将每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值，分别与预设条件中的PRB调度资源阈值、RSRP阈值和上行吞吐量阈值进行比较，判断是否存在满足预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值；若存在，则将与满足预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值相关的结果集合，确定为满足预设条件的聚类数据。
在一些实施例中，上行吞吐量阈值大于PRB调度资源阈值的3倍，且上行吞吐量阈值大于RSRP阈值的3倍。
在一些实施例中，从满足预设条件的聚类数据中确定出受扰站和受扰小区，包括：对于每两个结果集合组成的满足预设条件的聚类数据，将其中上行吞吐量平均值较小 的结果集合作为目标集合；将目标集合中的基站和小区分别作为受扰站和受扰小区。
在一些实施例中，调用MDA功能确定宏基站，包括：调用MDA功能从基站工参中抽取基站经纬度数据；基于基站经纬度数据，确定出与受扰站的距离小于距离阈值的基站作为宏基站。
在一些实施例中，根据受扰小区和宏基站确定施扰站，包括：调用MDA功能提取宏基站的RIM-RS(Remote Interference Management-Reference Signal，远程干扰管理参考信号)信息；其中，宏基站的RIM-RS信息包括：宏基站小区标识、宏基站波束标识、受干扰小区发送的干扰信息；基于RIM-RS信息对受扰小区进行关联分析，从宏基站中确定出施扰站；其中，施扰站的施扰站信息包括：施扰站标识、施扰小区标识、施扰波束标识。
在一些实施例中，调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略，包括：构建目标函数和相应的约束条件；其中，目标函数为受扰站下所有用户的吞吐量与施扰站下所有用户的吞吐量的加权值的最大化函数；利用构建的深度强化学习算法，根据目标函数和约束条件求解出干扰优化策略。
在一些实施例中，利用构建的深度强化学习算法，根据目标函数和约束条件求解出干扰优化策略，包括：配置动作集与状态集；
构建策略网络和动作评估网络；其中，策略网络和动作评估网络采用相同的三层网络结构；利用策略网络，根据动作集与状态集基于policy策略确定出策略；利用动作评估网络对策略收敛求解，得出干扰优化策略。
在一些实施例中，动作集包括：施扰站调整干扰的波束SSB方向、施扰站降低发射功率3dB、施扰站降低发射功率6dB、关闭施扰站产生干扰的时隙资源和关闭受扰站产生干扰的时隙资源；状态集包括：受扰站上报的性能数据和MR数据中所调度的PRB调度资源、RSRP和上行吞吐量。
在一些实施例中，在将干扰优化策略下发至施扰站和/或受扰站之后，还包括：采集受扰站和受扰小区的优化后指标数据，调用MDA功能，基于优化后指标数据判断是否存在干扰，以实现优化反馈。
根据本公开的另一个方面，提供一种干扰优化装置，包括：
数据采集模块，用于采集指标数据；判断受扰模块，用于调用MDA功能构造的聚类算法，基于所述指标数据判断是否存在干扰；确定受扰站模块，用于在存在干扰 的情况下确定出受扰站和受扰小区；确定施扰站模块，用于在存在干扰的情况下，调用MDA功能确定宏基站，根据受扰小区和宏基站确定施扰站；生成策略模块，用于调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略；下发模块，用于将干扰优化策略下发至施扰站和/或受扰站，以实现干扰优化。
在一些实施例中，指标数据包括：终端上报基站的MR数据和基站上报的基站性能数据；其中，MR数据的基础字段包括：基站标识和基站小区标识；MR数据的专属字段包括：RSRP；基站性能数据的基础字段包括：基站标识和基站小区标识；基站性能数据的专属字段包括：上行吞吐量、上行PRB调度资源；以及，在采集指标数据后，还包括：将MR数据和基站性能数据通过相同的基础字段进行关联存储。
在一些实施例中，判断受扰模块调用MDA功能构造的聚类算法，基于指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区，包括：调用MDA功能获取指标数据，抽取指标数据中的PRB调度资源、RSRP和上行吞吐量字段以组成三维数组；调用MDA功能，使用k-means算法对三维数组进行聚类分析，得到聚类结果；判断聚类结果中是否存在满足预设条件的聚类数据；若存在，则确定存在干扰；以及在确定存在干扰后，从满足预设条件的聚类数据中确定出受扰站和受扰小区。
在一些实施例中，判断受扰模块调用MDA功能，使用k-means算法对三维数组进行聚类分析，得到聚类结果，包括：以三维数组的个数作为观察值p，将p个三维数组分为k个初始集合；其中，k≤p；以最小化集群内平方和为目标进行聚类得到k个结果集合，以作为聚类结果。
在一些实施例中，判断受扰模块判断聚类结果中是否存在满足预设条件的聚类数据，包括：计算k个结果集合中每个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值；根据每两个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值，确定每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值；将每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值，分别与预设条件中的PRB调度资源阈值、RSRP阈值和上行吞吐量阈值进行比较，判断是否存在满足预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值；若存在，则将与满足预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值相关的结果集合，确定为满足预设条件的聚类数据。
在一些实施例中，上行吞吐量阈值大于PRB调度资源阈值的3倍，且上行吞吐量阈值大于RSRP阈值的3倍。
在一些实施例中，确定受扰站模块从满足预设条件的聚类数据中确定出受扰站和受扰小区，包括：对于每两个结果集合组成的满足预设条件的聚类数据，将其中上行吞吐量平均值较小的结果集合作为目标集合；将目标集合中的基站和小区分别作为受扰站和受扰小区。
在一些实施例中，确定施扰站模块调用MDA功能确定宏基站，包括：调用MDA功能从基站工参中抽取基站经纬度数据；基于基站经纬度数据，确定出与受扰站的距离小于距离阈值的基站作为宏基站。
在一些实施例中，确定施扰站模块根据受扰小区和宏基站确定施扰站，包括：调用MDA功能提取宏基站的远程干扰管理参考信号RIM-RS信息；其中，宏基站的RIM-RS信息包括：宏基站小区标识、宏基站波束标识、受干扰小区发送的干扰信息；基于RIM-RS信息对受扰小区进行关联分析，从宏基站中确定出施扰站；其中，施扰站的施扰站信息包括：施扰站标识、施扰小区标识、施扰波束标识。
在一些实施例中，生成策略模块调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略，包括：构建目标函数和相应的约束条件；其中，目标函数为受扰站下所有用户的吞吐量与施扰站下所有用户的吞吐量的加权值的最大化函数；利用构建的深度强化学习算法，根据目标函数和约束条件求解出干扰优化策略。
在一些实施例中，生成策略模块利用构建的深度强化学习算法，根据目标函数和约束条件求解出干扰优化策略，包括：配置动作集与状态集；
构建策略网络和动作评估网络；其中，策略网络和动作评估网络采用相同的三层网络结构；利用策略网络，根据动作集与状态集基于policy策略确定出策略；利用动作评估网络对策略收敛求解，得出干扰优化策略。
在一些实施例中，动作集包括：施扰站调整干扰的波束SSB方向、施扰站降低发射功率3dB、施扰站降低发射功率6dB、关闭施扰站产生干扰的时隙资源和关闭受扰站产生干扰的时隙资源；状态集包括：受扰站上报的性能数据和MR数据中所调度的PRB调度资源、RSRP和上行吞吐量。
在一些实施例中，下发模块还用于：在将干扰优化策略下发至施扰站和/或受扰站之后，还包括：采集受扰站和受扰小区的优化后指标数据，调用MDA功能，基于优 化后指标数据判断是否存在干扰，以实现优化反馈。
根据本公开的又一个方面，提供一种非易失性计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现上述的干扰优化方法。
根据本公开的再一个方面，提供一种电子设备，包括：处理器；以及存储器，用于存储所述处理器的可执行指令；其中，所述处理器配置为经由执行所述可执行指令来执行上述的干扰优化方法。
根据本公开的再一个方面，提供一种计算机程序，包括：指令，所述指令当由处理器执行时使所述处理器执行上述任一个实施例的干扰优化方法。
根据本公开的再一个方面，提供一种计算机程序产品，包括指令，所述指令当由处理器执行时使所述处理器执行上述任一个实施例的干扰优化方法。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，并不能限制本公开。
附图说明
此处的附图被并入说明书中并构成本说明书的一部分，示出了符合本公开的实施例，并与说明书一起用于解释本公开的原理。显而易见地，下面描述中的附图仅仅是本公开的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1示出了可以应用本公开一些实施例的干扰优化方法的示例性系统架构的示意图；
图2示出了本公开一些实施例的干扰优化方法的流程图；
图3示出了本公开一些实施例的判断是否存在干扰的方法流程图；
图4示出了本公开一些实施例的确定受扰站的方法流程图；
图5示出了本公开一些实施例的在存在干扰的情况下确定施扰站的方法流程图；
图6示出了本公开一些实施例的生成干扰优化策略的方法流程图；
图7示出了本公开一些实施例的用于实现干扰优化的装置示意图；
图8示出了本公开一些实施例的干扰优化装置的框图；和
图9示出了本公开一些实施例中一种干扰优化计算机设备的结构框图。
具体实施方式
现在将参考附图更全面地描述示例实施方式。然而，示例实施方式能够以多种形式实施，且不应被理解为限于在此阐述的范例；相反，提供这些实施方式使得本公开将更加全面和完整，并将示例实施方式的构思全面地传达给本领域的技术人员。所描述的特征、结构或特性可以以任何合适的方式结合在一个或更多实施方式中。
此外，附图仅为本公开的示意性图解，并非一定是按比例绘制。图中相同的附图标记表示相同或类似的部分，因而将省略对它们的重复描述。附图中所示的一些方框图是功能实体，不一定必须与物理或逻辑上独立的实体相对应。可以采用软件形式来实现这些功能实体，或在一个或多个硬件模块或集成电路中实现这些功能实体，或在不同网络和/或处理器装置和/或微控制器装置中实现这些功能实体。
此外，术语“第一”、“第二”仅用于描述目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括一个或者更多个该特征。在本公开的描述中，“多个”的含义是至少两个，例如两个，三个等，除非另有明确具体的限定。
在5G应用中，针对为满足5G行业用户对大上行能力的要求而采用的TDD频段灵活双工组网配置方案而存在的交叉干扰问题，业界已分别从施扰侧和受扰侧进行了干扰优化方案研究，分析不同负载或覆盖下干扰优化策略的选择。
目前3GPP TR28.809协议中确定了网管系统中的MDAS(Management Data Analytics service,数据分析管理服务)在管理循环中的角色及作用，数据分析管理服务可基于机器学习模型识别网络性能劣化问题，并精准识别问题根因，优化策略解决问题，确定使用者和提供者之间的交互反馈机制，相关场景的数据输入信息和输出属性。但目前标准和实现上还有如下问题：1)缺乏网管侧的全控管理及具体的分析方法。目前针对TDD频段灵活双工组网配置方案的交叉干扰问题解决方法是基站侧去主动识别干扰问题、受扰站发送信号到施扰站确定干扰源，再手动获取网管配置干扰优化策略，一方面增加了基站侧的计算复杂度，方法缺乏灵活性。另一方面，受扰站和施扰站之间的信息传输会存在信息丢失，降低干扰问题解决的时效性，间歇性影响用户的数据质量；2)进一步，网管侧缺乏针对该场景的具体数据处理机制：实现网管的全控机制需要借助MDA(Management Data Analytics,管理数据分析)功能对干扰问题进行问题识别以得出干扰优化策略，然而目前标准中缺乏针对该场景的具体分析过程，无法指引RAN(Radio Access Network，无线接入网)侧集中数据采集单元进行相应指标数据采集及数据分析方向，也无法确定输出属性的需求，影响了上述场景下干扰 优化智能化解决方案的实现。
目前的干扰优化方法中基站侧的计算复杂度较高，受扰站和施扰站之间的信息传输常常会存在信息丢失，降低了干扰问题解决的时效性较差，从而影响了用户的数据质量。
基于此，本公开的示例性实施方式提供了一种干扰优化方法，以用于至少解决上述技术问题中的一个或者全部。
本公开的目的在于提供一种干扰优化方法、装置、电子设备及存储介质，以提供一种由网管系统执行的干扰优化方法，从网管全控角度，引入数据分析管理服务MDAS机制，能够确认干扰现象、确定受扰站、确定施扰站、生成干扰优化策略并下发至受扰站和/或施扰站，进而实现网管侧主导的干扰优化的全闭环流程。
图1示出了可以应用本公开一些实施例的干扰优化方法的示例性系统架构的示意图；如图1所示：
该系统架构可以包括服务器101、网络102和客户端103。网络102用以在客户端103和服务器101之间提供通信链路的介质。网络102可以包括各种连接类型，例如有线、无线通信链路或者光纤电缆等等。
服务器101可以为网管控制提供服务，可以是提供各种服务的服务器，例如提供接收终端(客户端103)上报的指标数据、根据指标数据判断是否存在干扰现象、在存在干扰现象时确定出受扰站和施扰站、生成干扰优化策略并下发等功能服务的后台管理服务器。即：设置在网管系统的后台管理服务器可以根据客户端103上报的指标数据确认是否存在干扰现象、确定受扰站和施扰站、生成干扰优化策略并下发，从而实现网管侧主导的干扰优化全闭环流程，提升客户端103的数据质量。
客户端103可以是手机、游戏主机、平板电脑、电子书阅读器、智能眼镜、智能家居设备、AR(Augmented Reality，增强现实)设备、VR(Virtual Reality，虚拟现实)设备等移动终端，或者，客户端103也可以是个人计算机，比如膝上型便携计算机和台式计算机等等。
在一些可选的实施例中，服务器101可以在接收到客户端103上报的指标数据后，调用管理数据分析MDA功能构造的聚类算法根据指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区；还可以调用MDA功能确定宏基站，以根据受扰小区和宏基站确定施扰站；再调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略，以将干扰优化策略下发 至施扰站和/或受扰站，进而实现干扰优化。服务器101还可以在下发干扰优化策略后采集受扰站和受扰小区下客户端103上报的优化后指标数据，以实现优化反馈。
应该理解，图1中的客户端、网络和服务器的数目仅仅是示意性的，服务器101可以是一个实体的服务器，还可以为多个服务器组成的服务器集群，还可以是云端服务器，根据实际需要，可以具有任意数目的客户端、网络和服务器。
下面，将结合附图及实施例对本公开示例实施例中的干扰优化方法的各个步骤进行更详细的说明。
图2示出了本公开一些实施例的干扰优化方法的流程图。本公开实施例提供的方法可以由如图1所示的服务器或客户端中执行，但本公开并不限定于此。
在下面的举例说明中，以服务器集群101为执行主体进行示例说明。
如图2所示，本公开实施例提供的干扰优化方法可以由网管系统执行，包括以下步骤：
步骤S201，采集指标数据。
在一些实施例中，网管系统可以采集并监控一定时间段内终端上报基站的MR(Measurement Report，测量报告)数据、基站上报的性能数据，指标数据可以包括：终端上报基站的测量报告MR数据和基站上报的基站性能数据；其中，MR数据的基础字段包括：基站标识和基站小区标识；MR数据的专属字段包括：RSRP(Reference Signal Received Power，参考信号接收功率)；基站性能数据的基础字段包括：基站标识和基站小区标识；基站性能数据的专属字段包括：上行吞吐量、上行物理资源块PRB(Physical Resource Block，物理资源块)调度资源；以及，在采集指标数据后，还可以将MR数据和基站性能数据通过相同的基础字段进行关联存储。
步骤S203，调用管理数据分析MDA功能构造的聚类算法，基于指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区。步骤S205，在存在干扰的情况下，调用MDA功能确定宏基站，根据受扰小区和宏基站确定施扰站。步骤S207，调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略。步骤S209，将干扰优化策略下发至施扰站和/或受扰站，以实现干扰优化。
本申请提出一种有网管系统执行的基于网管数据分析处理的干扰优化方法，主要思想为：面向5G大上行用户部署的TDD灵活双工组网场景，网管系统实时采集终端上报的MR数据及基站上报的基站性能等指标数据，调用MDA功能构造的聚类算法 分析PRB、RSRP和上行吞吐量数据以主动识别干扰现象的存在，并找出受扰站；以及当发现存在干扰时，可以分析受扰站周边宏基站的RIM-RS信号信息以定位发现施扰站；再调用MDA构造的强化学习算法分析生成干扰优化策略，以将干扰优化策略的配置信息下发相应施扰站或受扰站(以)完成干扰优化处理。网管系统利用管理数据分析MDA功能实现了干扰问题识别、受扰站和干扰源确定、干扰优化策略生成的全控闭环流程。
在一些实施例中，调用管理数据分析MDA功能构造的聚类算法，基于指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区，包括：调用MDA功能获取指标数据，抽取指标数据中的PRB调度资源、RSRP和上行吞吐量字段以组成三维数组；调用MDA功能，使用k-means算法对三维数组进行聚类分析，得到聚类结果；判断聚类结果中是否存在满足预设条件的聚类数据；若存在，则确定存在干扰；以及在确定存在干扰后，从满足预设条件的聚类数据中确定出受扰站和受扰小区。
本公开的实施例所提供的干扰优化方法，能够从网管全控角度，通过引入数据分析管理服务MDAS机制来确认干扰现象、确定受扰站、确定施扰站、生成干扰优化策略并下发至受扰站和/或施扰站，从而实现网管侧主导的干扰优化的全闭环流程。
图3示出了本公开一些实施例的判断是否存在干扰的方法流程图，如图3所示，包括：
步骤301，调用MDA功能获取指标数据，抽取指标数据中的PRB调度资源、RSRP和上行吞吐量字段以组成三维数组；
步骤303，调用MDA功能，使用k-means算法对三维数组进行聚类分析，得到聚类结果；
步骤305，判断聚类结果中是否存在满足预设条件的聚类数据；若是，则在步骤307中判断为存在干扰；若否，则在步骤309中判断为不存在干扰。
进一步地，在一些实施例中，网管系统可以调用MDA功能，使用k-means算法对三维数组进行聚类分析，得到聚类结果，包括：以三维数组的个数作为观察值p，将p个三维数组分为k个初始集合；其中，k≤p；以最小化集群内平方和为目标进行聚类得到k个结果集合，以作为聚类结果。
网管的MDA功能可以用于分析数据以识别干扰是否存在。具体地，可以基于k-means算法对三维数组进行聚类分析，其中，数组可以定义为(X 1，X 2，…，X p)，根据数组的个数定义观察值为p，每组数据可以包括前述PRB、RSRP、上行吞吐量字 段；再将观察目标p个数组聚类为k(其中k<＝p)个结果集合，k个结果集合组成的集合S＝{S 1，S 2，…，S k}，再以最小化集群内平方和为目标进行聚类。
其中目标可以定义为：
Figure PCTCN2022103772-appb-000001
μ i是每组的聚类中心点。
聚类结束后，可以在每组数据后面添加一个聚类分组标识(如：1、2、3……k)，以作为最终的聚类结果进行输出。
在一些实施例中，判断聚类结果中是否存在满足预设条件的聚类数据，包括：计算k个结果集合中每个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值；根据每两个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值，确定每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值；将每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值，分别与预设条件中的PRB调度资源阈值、RSRP阈值和上行吞吐量阈值进行比较，判断是否存在满足预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值；若存在，则将与满足预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值相关的结果集合，确定为满足预设条件的聚类数据。进一步地，在一些实施例中，上行吞吐量阈值大于PRB调度资源阈值的3倍，且上行吞吐量阈值大于RSRP阈值的3倍。
可以计算聚类结果中每个结果集合的PRB、RSRP、吞吐量的平均值，并进行两两对比，筛选是否有符合以下条件的数据：
||Avg prb_i-Avg prb_j||<A&||Avg rsrp_i-Avg rsrp_j||<B&||Avg thp_i-Avg thp_j||>C
可以先确定任意两个结果集合S i和S j，Avg prb_i与Avg prb_j分别表示这两个结果集合的PRB调度资源平均值；Avg rsrp_i与Avg rsrp_j分别表示这两个结果集合的RSRP平均值；Avg thp_i与Avg thp_j分别表示这两个结果集合的上行吞吐量平均值；其中A表示PRB调度资源阈值，B表示RSRP阈值，C表示上行吞吐量阈值，A和B的取值可以参照实际应用来确定，A、B和C之间的关系应保持：C为A的3倍以上，以及C为B的3倍以上。
在一些实施例中，从满足预设条件的聚类数据中确定出受扰站和受扰小区，包括：对于每两个结果集合组成的满足预设条件的聚类数据，将其中上行吞吐量平均值较小 的结果集合作为目标集合；将目标集合中的基站和小区分别作为受扰站和受扰小区。
当存在两个结果集合满足上述不等式关系，则可以认为这两个结果集合所涉及的终端存在交叉链路干扰问题，可以提取这两个集合中上行吞吐量平均值较低的集合，将集合中相应的基站和小区分别作为受扰站和受扰小区进行输出。
图4示出了本公开一些实施例的确定受扰站的方法流程图，如图4所示，包括：
步骤401，计算k个结果集合中每个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值；
步骤403，根据每两个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值，确定每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值；
步骤405，将每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值，分别与预设条件中的PRB调度资源阈值、RSRP阈值和上行吞吐量阈值进行比较；
步骤407，当存在满足预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值时，将与满足预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值相关的结果集合，确定为满足预设条件的聚类数据；
步骤409，对于每两个结果集合组成的满足预设条件的聚类数据，将其中上行吞吐量平均值较小的结果集合作为目标集合；
步骤411，将目标集合中的基站和小区分别作为受扰站和受扰小区。
在一些实施例中，调用MDA功能确定宏基站，包括：调用MDA功能从基站工参中抽取基站经纬度数据；基于基站经纬度数据，确定出与受扰站的距离小于距离阈值的基站作为宏基站。
网管系统可以抽取基站工参中的经纬度数据，进而获取受扰站的经纬度数据以及受扰站周边范围内的第一宏基站的经纬度数据，再通过受扰站与第一宏基站之间的距离，找出与受扰站的距离S在阈值范围D内的第一宏基站，以作为上述实施例中的宏基站。
根据两个基站的经纬度(Lat1，Lng1)(Lat2，Lat2)计算二者之间距离S的计算方式如下：
Figure PCTCN2022103772-appb-000002
Figure PCTCN2022103772-appb-000003
在一些实施例中，根据受扰小区和宏基站确定施扰站，包括：调用MDA功能提取宏基站的远程干扰管理参考信号RIM-RS信息；其中，宏基站的RIM-RS信息包括：宏基站小区标识、宏基站波束标识、受干扰小区发送的干扰信息；基于RIM-RS信息对受扰小区进行关联分析，从宏基站中确定出施扰站；其中，施扰站的施扰站信息包括：施扰站标识、施扰小区标识、施扰波束标识。
图5示出了本公开一些实施例的在存在干扰的情况下确定施扰站的方法流程图，如图5所示，包括：
步骤501，调用MDA功能从基站工参中抽取基站经纬度数据；在一些实际应用中，也可以从其他记录了经纬度信息的数据表中抽取基站经纬度数据；
步骤503，根据基站经纬度数据确定受扰站经纬度数据，以及受扰站周边的第一宏基站的经纬度数据；
步骤505，计算受扰站与第一宏基站之间的距离S，找出小于距离阈值D的S所对应的第一宏基站，作为宏基站；
步骤507，调用MDA功能提取宏基站的远程干扰管理参考信号RIM-RS信息；其中，宏基站的RIM-RS信息包括：宏基站小区标识、宏基站波束标识、受干扰小区发送的干扰信息；
步骤509，基于RIM-RS信息对受扰小区进行关联分析，从宏基站中确定出施扰站；其中，施扰站的施扰站信息包括：施扰站标识、施扰小区标识、施扰波束标识。
在一些实施例中，调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略，包括：构建目标函数和相应的约束条件；其中，目标函数为受扰站下所有用户的吞吐量与施扰站下所有用户的吞吐量的加权值的最大化函数；利用构建的深度强化学习算法，根据目标函数和约束条件求解出干扰优化策略。
引入了MDAS的网管系统中的网管MDA功能，可以通过建立好的算法模型实现干扰优化策略的实时在线生成。
对于目标函数，可以设置目标函数的意义为使得单位时间内受扰站下所有用户的吞吐量与施扰站下所有用户的吞吐量加权值最大；目标函数可以如下设置：
Figure PCTCN2022103772-appb-000004
其中，Thp(t)表示单位时间内的受扰站下所有用户的吞吐量与施扰站下所有用户的吞吐量加权值；N表示受扰站下所有用户的总数，M表示施扰站下所有用户的总数。
对于约束条件，可以设置使约束条件的约束效果为：以尽最大可能保障受扰站用户的QoS(Quality of Service，服务质量)，同时降低对施扰站用户性能影响；约束条件可以如下设置：
Figure PCTCN2022103772-appb-000005
基于前述步骤进行聚类后的聚类结果，可以先确定任意两个结果集合S i和S j，约束条件中的Avg prb_i与Avg prb_j分别表示这两个结果集合的PRB调度资源平均值；Avg rsrp_i与Avg rsrp_j分别表示这两个结果集合的RSRP平均值；Avg thp_i与Avg thp_j分别表示这两个结果集合的上行吞吐量平均值；其中A表示PRB调度资源阈值，B表示RSRP阈值，C表示上行吞吐量阈值，A和B的取值可以参照实际应用来确定，A、B和C之间的关系应保持C为A的3倍以上，以及C为B的3倍以上。则约束条件中第一个条件的设置可以使得前述步骤进行聚类后聚类结果中的各结果集合prb、吞吐量、RSRP均在相应阈值之内；
约束条件中第二个条件的设置，可以使受扰站下所有用户的吞吐总量不能低于μ 0，
约束条件中第二个条件的设置，可以使施扰站下所有用户的吞吐总量低于q的概率小于τ，其中q值的确定可以按照施扰站下用户负荷量m来确定，确定关系成负相关：
Figure PCTCN2022103772-appb-000006
在一些实施例中，利用构建的深度强化学习算法，根据目标函数和约束条件求解出干扰优化策略，包括：配置动作集与状态集；
构建策略网络和动作评估网络；其中，策略网络和动作评估网络采用相同的三层网络结构；利用策略网络，根据动作集与状态集基于policy策略确定出策略；利用动作评估网络对策略收敛求解，得出干扰优化策略。
在一些实际应用中，所述三层网络结构可以包括：设置神经元+rule函数的输入 层，设置神经元+relu函数的中间层和设置神经元+sigmoid+softmax两个函数的输出层。
在一些实施例中，动作集包括：施扰站调整干扰的波束SSB方向、施扰站降低发射功率3dB、施扰站降低发射功率6dB、关闭施扰站产生干扰的时隙资源和关闭受扰站产生干扰的时隙资源；动作集中的数据可以认为是5个属性的转移动作。
状态集包括：受扰站上报的性能数据和MR数据中所调度的PRB调度资源、RSRP和上行吞吐量。
使用本申请中的策略生成方法，可以基于前序步骤的输出结果进行策略生成，可以包括构建马尔科夫元组用户状态和当前时刻到下一时刻的动作，由状态和动作构建系统奖励函数，构造深度强化学习模型确定生成最优的干扰优化策略。
图6示出了本公开一些实施例的生成干扰优化策略的方法流程图，如图6所示，包括：
步骤S601，构建目标函数和相应的约束条件；
步骤S603，配置动作集与状态集，以及构建策略网络和动作评估网络；
步骤S605，利用策略网络，根据动作集与状态集基于policy策略确定出策略；
步骤S607，利用动作评估网络对策略收敛求解，得出干扰优化策略。
在一些实际应用中，可以通过MDA功能构造的强化学习算法得到干扰优化策略的配置信息，配置信息中可以指示与动作集中的5个转移动作相关的调整策略；将干扰优化策略下发至干扰站和/或受扰站后，可以使干扰站和/或受扰站基于策略中的配置信息做出具体的调整，即：可以通过MDA功能构造的强化学习算法得到具体的输出属性，用于使干扰站和/或受扰站执行进而解决干扰问题。
在一些实施例中，在将干扰优化策略下发至施扰站和/或受扰站之后，还包括：采集受扰站和受扰小区的优化后指标数据，调用MDA功能，基于优化后指标数据判断是否存在干扰，以实现优化反馈。
使用本申请中干扰优化方法，可以利用网管设备将MDA功能生成的干扰配置信息下发至干扰站和/或施扰站生效。网管实时采集优化后的受扰侧数据，确定受扰侧的干扰信息是否解决，实时迭代进行MDA处理分析。本申请中的方法具备验证反馈功能，MDA实时在线更新模型参数及策略的过程，迭代优化策略，进而产生更好的优化效果。
图7示出了本公开一些实施例的用于实现干扰优化的装置示意图，如图7所示， 包括：
RAN域网管设备，受扰站，施扰站；
在RAN网管设备侧，可以利用引入的MDA功能实现数据采集分类、干扰问题的识别分析、干扰站的确定分析、干扰优化策略的生成分析等功能，并在生成干扰优化策略后，将策略下发至干扰站和/或施扰站生效，实时解决干扰问题。在一些实际应用中，在干扰站和/或施扰站执行干扰优化策略后，网管设备侧还可以采集受扰侧的优化后数据，进行分析以及时反馈优化效果，及时更新学习模型以调整优化方案，达到更好的优化效果。
在一些实际应用中，本申请具备数据采集及配置信息下发交互机制，可以实时更新MDA功能中模型的建立，优化输出，并且设备间的交互流程与现有网络管理架构相融合，便于部署实施。
根据本申请提供的干扰优化方法，一方面可以针对TDD频段灵活双工组网配置场景的干扰优化问题，通过在网管侧引入MDAS机制，使网管侧能够采集到该场景下用于干扰优化的输入数据，能够主动进行模型构建，以及能够通过模型得到用于优化的输出属性；其中，网管侧的RAN集中数据采集单元可以实现指标数据采集，引入的MDA功能可以实现模型构建及数据分析，进而能够从网管全控角度实现干扰问题的主动识别、干扰源的确定、干扰优化策略的生成这种全控闭环流程，进而可以保障该场景下干扰问题识别的实时性、生成干扰策略的高效性，为有大上行需求的垂直行业用户的性能提供可靠保证。
另一方面，从实现角度，本申请提出的网管系统增加了该场景的MDA功能，能够从网管侧实现数据的采集输入、分析机制、输出属性，运营商可以有针对性地实现数据采集及有目的性地进行模型算法的开发，为运营商面向垂直行业对上行要求较高的行业部署组网优化升级提供基础支撑。
本申请提供的干扰优化方法还能够减少基站的计算复杂度，从网管角度进行数据处理分析并自动下发干扰优化策略配置信息生效，减少了运维人员手动配置运维的负担，提高了运维效率。
需要注意的是，上述附图仅是根据本公开示例性实施例的方法所包括的处理的示意性说明，而不是限制目的。易于理解，上述附图所示的处理并不表明或限制这些处理的时间顺序。另外，也易于理解，这些处理可以是例如在多个模块中同步或异步执行的。
图8示出了本公开一些实施例的干扰优化装置800的框图；如图8所示，包括：
数据采集模块801，用于采集指标数据；
判断受扰模块802，用于调用管理数据分析MDA功能构造的聚类算法，基于所述指标数据判断是否存在干扰；
确定受扰站模块803，用于在存在干扰的情况下确定出受扰站和受扰小区；
确定施扰站模块804，用于在存在干扰的情况下，调用MDA功能确定宏基站，根据受扰小区和宏基站确定施扰站；
生成策略模块805，用于调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略；
下发模块806，用于将干扰优化策略下发至施扰站和/或受扰站，以实现干扰优化。
所属技术领域的技术人员能够理解，本公开的各个方面可以实现为系统、方法或程序产品。因此，本公开的各个方面可以具体实现为以下形式，即：完全的硬件实施方式、完全的软件实施方式(包括固件、微代码等)，或硬件和软件方面结合的实施方式，这里可以统称为“电路”、“模块”或“系统”。
图9示出本公开一些实施例中一种干扰优化计算机设备的结构框图。需要说明的是，图示出的电子设备仅仅是一个示例，不应对本公开实施例的功能和使用范围带来任何限制。
下面参照图9来描述根据本公开的这种实施方式的电子设备900。图9显示的电子设备900仅仅是一个示例，不应对本公开实施例的功能和使用范围带来任何限制。
如图9所示，电子设备900以通用计算设备的形式表现。电子设备900的组件可以包括但不限于：上述至少一个处理单元910、上述至少一个存储单元920、连接不同系统组件(包括存储单元920和处理单元910)的总线930。
其中，所述存储单元存储有程序代码，所述程序代码可以被所述处理单元910执行，使得所述处理单元910执行本说明书上述“示例性方法”部分中描述的根据本公开各种示例性实施方式的步骤。例如，所述处理单元910可以执行如图2中所示的步骤S201，采集指标数据；步骤S203，调用管理数据分析MDA功能构造的聚类算法，基于指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区。步骤S205，在存在干扰的情况下，调用MDA功能确定宏基站，根据受扰小区和宏基站确定施扰站。步骤S207，调用MDA功能构造的强化学习算法，基于受扰站的受扰站信息和施扰站的施扰站信息生成干扰优化策略。步骤S209，将干扰优化策略下发至 施扰站和/或受扰站，以实现干扰优化。
存储单元920可以包括易失性存储单元形式的可读介质，例如随机存取存储单元(RAM)9201和/或高速缓存存储单元9202，还可以进一步包括只读存储单元(ROM)9203。
存储单元920还可以包括具有一组(至少一个)程序模块9205的程序/实用工具9204，这样的程序模块9205包括但不限于：操作系统、一个或者多个应用程序、其它程序模块以及程序数据，这些示例中的每一个或某种组合中可能包括网络环境的实现。
总线930可以为表示几类总线结构中的一种或多种，包括存储单元总线或者存储单元控制器、外围总线、图形加速端口、处理单元或者使用多种总线结构中的任意总线结构的局域总线。
电子设备900也可以与一个或多个外部设备干扰优化装置800(例如键盘、指向设备、蓝牙设备等)通信，还可与一个或者多个使得用户能与该电子设备900交互的设备通信，和/或与使得该电子设备900能与一个或多个其它计算设备进行通信的任何设备(例如路由器、调制解调器等等)通信。这种通信可以通过输入/输出(I/O)接口950进行。并且，电子设备900还可以通过网络适配器960与一个或者多个网络(例如局域网(LAN)，广域网(WAN)和/或公共网络，例如因特网)通信。如图所示，网络适配器960通过总线930与电子设备900的其它模块通信。应当明白，尽管图中未示出，可以结合电子设备900使用其它硬件和/或软件模块，包括但不限于：微代码、设备驱动器、冗余处理单元、外部磁盘驱动阵列、RAID系统、磁带驱动器以及数据备份存储系统等。
通过以上的实施方式的描述，本领域的技术人员易于理解，这里描述的示例实施方式可以通过软件实现，也可以通过软件结合必要的硬件的方式来实现。因此，根据本公开实施方式的技术方案可以以软件产品的形式体现出来，该软件产品可以存储在一个非易失性存储介质(可以是CD-ROM，U盘，移动硬盘等)中或网络上，包括若干指令以使得一台计算设备(可以是个人计算机、服务器、终端装置、或者网络设备等)执行根据本公开实施方式的方法。
在本公开的示例性实施例中，还提供了一种计算机可读存储介质，其上存储有能够实现本说明书上述方法的程序产品。在一些可能的实施方式中，本公开的各个方面还可以实现为一种程序产品的形式，其包括程序代码，当所述程序产品在终端设备上运行时，所述程序代码用于使所述终端设备执行本说明书上述“示例性方法”部分中 描述的根据本公开各种示例性实施方式的步骤。
在本公开的示例性实施例中，还提供了一种计算机程序，包括：指令，所述指令当由处理器执行时使所述处理器执行上述任一个实施例的干扰优化方法。
在本公开的示例性实施例中，还提供了一种计算机程序产品，包括指令，所述指令当由处理器执行时使所述处理器执行上述任一个实施例的干扰优化方法。
根据本公开实施方式的用于实现上述方法的程序产品，其可以采用便携式紧凑盘只读存储器(CD-ROM)并包括程序代码，并可以在终端设备，例如个人电脑上运行。然而，本公开的程序产品不限于此，在本文件中，可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。
所述程序产品可以采用一个或多个可读介质的任意组合。可读介质可以是可读信号介质或者可读存储介质。可读存储介质例如可以为但不限于电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。可读存储介质的更具体的例子(非穷举的列表)包括：具有一个或多个导线的电连接、便携式盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、光纤、便携式紧凑盘只读存储器(CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。
计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了可读程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。可读信号介质还可以是可读存储介质以外的任何可读介质，该可读介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。
可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于无线、有线、光缆、RF等等，或者上述的任意合适的组合。
可以以一种或多种程序设计语言的任意组合来编写用于执行本公开操作的程序代码，所述程序设计语言包括面向对象的程序设计语言—诸如Java、C++等，还包括常规的过程式程序设计语言—诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算设备上执行、部分地在用户设备上执行、作为一个独立的软件包执行、部分在用户计算设备上部分在远程计算设备上执行、或者完全在远程计算设备或服务器上执行。在涉及远程计算设备的情形中，远程计算设备可以通过任意种类的网 络，包括局域网(LAN)或广域网(WAN)，连接到用户计算设备，或者，可以连接到外部计算设备(例如利用因特网服务提供商来通过因特网连接)。
应当注意，尽管在上文详细描述中提及了用于动作执行的设备的若干模块或者单元，但是这种划分并非强制性的。实际上，根据本公开的实施方式，上文描述的两个或更多模块或者单元的特征和功能可以在一个模块或者单元中具体化。反之，上文描述的一个模块或者单元的特征和功能可以进一步划分为由多个模块或者单元来具体化。
此外，尽管在附图中以特定顺序描述了本公开中方法的各个步骤，但是，这并非要求或者暗示必须按照该特定顺序来执行这些步骤，或是必须执行全部所示的步骤才能实现期望的结果。附加的或备选的，可以省略某些步骤，将多个步骤合并为一个步骤执行，以及/或者将一个步骤分解为多个步骤执行等。
通过以上的实施方式的描述，本领域的技术人员易于理解，这里描述的示例实施方式可以通过软件实现，也可以通过软件结合必要的硬件的方式来实现。因此，根据本公开实施方式的技术方案可以以软件产品的形式体现出来，该软件产品可以存储在一个非易失性存储介质(可以是CD-ROM，U盘，移动硬盘等)中或网络上，包括若干指令以使得一台计算设备(可以是个人计算机、服务器、移动终端、或者网络设备等)执行根据本公开实施方式的方法。
本领域技术人员在考虑说明书及实践这里公开的内容后，将容易想到本公开的其它实施方案。本申请旨在涵盖本公开的任何变型、用途或者适应性变化，这些变型、用途或者适应性变化遵循本公开的一般性原理并包括本公开未公开的本技术领域中的公知常识或惯用技术手段。说明书和实施例仅被视为示例性的，本公开的真正范围和精神由所附的权利要求指出。

Claims (20)

  1. 一种干扰优化方法，所述干扰优化方法由网管系统执行，包括：
    采集指标数据；
    调用管理数据分析MDA功能构造的聚类算法，基于所述指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区；
    在存在干扰的情况下，调用所述MDA功能确定宏基站，根据所述受扰小区和所述宏基站确定施扰站；
    调用所述MDA功能构造的强化学习算法，基于所述受扰站的受扰站信息和所述施扰站的施扰站信息生成干扰优化策略；
    将所述干扰优化策略下发至所述施扰站或所述受扰站中的至少一个，以实现干扰优化。
  2. 根据权利要求1所述的干扰优化方法，其中，所述指标数据包括终端上报基站的测量报告MR数据和基站上报的基站性能数据，所述MR数据的基础字段包括基站标识和基站小区标识，所述MR数据的专属字段包括参考信号接收功率RSRP，所述基站性能数据的基础字段包括基站标识和基站小区标识，所述基站性能数据的专属字段包括上行吞吐量、上行物理资源块PRB调度资源。
  3. 根据权利要求1所述的干扰优化方法，在采集所述指标数据后，所述干扰优化方法还包括：
    将所述MR数据和所述基站性能数据通过相同的基础字段进行关联存储。
  4. 根据权利要求3所述的干扰优化方法，其中，调用管理数据分析MDA功能构造的聚类算法，基于所述指标数据判断是否存在干扰，并在存在干扰的情况下确定出受扰站和受扰小区，包括：
    调用所述MDA功能获取所述指标数据，抽取所述指标数据中的PRB调度资源、RSRP和上行吞吐量字段以组成三维数组；
    调用所述MDA功能，对所述三维数组进行聚类分析，得到聚类结果；
    判断所述聚类结果中是否存在满足预设条件的聚类数据；
    在存在满足预设条件的聚类数据的情况下，确定存在干扰；以及
    在确定存在干扰后，从满足预设条件的聚类数据中确定出所述受扰站和所述受扰小区。
  5. 根据权利要求4所述的干扰优化方法，其中，所述对所述三维数组进行聚类分析包括：
    使用k-means算法对所述三维数组进行聚类分析。
  6. 根据权利要求4所述的干扰优化方法，其中，调用所述MDA功能，使用k-means算法对所述三维数组进行聚类分析，得到聚类结果，包括：
    以所述三维数组的个数作为观察值p，将所述p个三维数组分为k个初始集合，k≤p；
    以最小化集群内平方和为目标进行聚类，得到k个结果集合，以作为所述聚类结果。
  7. 根据权利要求6所述的干扰优化方法，其中，判断所述聚类结果中是否存在满足预设条件的聚类数据，包括：
    计算所述k个结果集合中每个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值；
    根据每两个结果集合的PRB调度资源平均值、RSRP平均值和上行吞吐量平均值，确定所述每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值；
    将所述每两个结果集合的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值，分别与所述预设条件中的PRB调度资源阈值、RSRP阈值和上行吞吐量阈值进行比较，以判断是否存在满足所述预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值；
    在存在满足所述预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值的情况下，将与满足所述预设条件中不等式关系的PRB调度资源均值差值、RSRP均值差值和上行吞吐量均值差值相关的结果集合，确定为所述满足预设条件的聚类数据。
  8. 根据权利要求7所述的干扰优化方法，其中，所述上行吞吐量阈值大于所述PRB调度资源阈值的3倍，且所述上行吞吐量阈值大于所述RSRP阈值的3倍。
  9. 根据权利要求7所述的干扰优化方法，其中，从满足预设条件的聚类数据中确定出所述受扰站和所述受扰小区，包括：
    对于每两个结果集合组成的满足预设条件的聚类数据，将所述满足预设条件的聚类数据中上行吞吐量平均值较小的结果集合作为目标集合；
    将所述目标集合中的基站和小区分别作为所述受扰站和所述受扰小区。
  10. 根据权利要求1所述的干扰优化方法，其中，调用所述MDA功能确定宏基站，包括：
    调用所述MDA功能从基站工参中抽取基站经纬度数据；
    基于所述基站经纬度数据，将与所述受扰站的距离小于距离阈值的基站确定为宏基站。
  11. 根据权利要求1所述的干扰优化方法，其中，根据所述受扰小区和所述宏基站确定施扰站，包括：
    调用所述MDA功能提取所述宏基站的远程干扰管理参考信号RIM-RS信息，所述宏基站的RIM-RS信息包括宏基站小区标识、宏基站波束标识、受干扰小区发送的干扰信息；
    基于所述RIM-RS信息对所述受扰小区进行关联分析，从所述宏基站中确定出所述施扰站，所述施扰站的施扰站信息包括施扰站标识、施扰小区标识、施扰波束标识。
  12. 根据权利要求1所述的干扰优化方法，其中，调用所述MDA功能构造的强化学习算法，基于所述受扰站的受扰站信息和所述施扰站的施扰站信息生成干扰优化策略，包括：
    构建目标函数和所述目标函数相应的约束条件，所述目标函数为受扰站下所有用户的吞吐量与施扰站下所有用户的吞吐量的加权值的最大化函数；
    利用构建的深度强化学习算法，根据所述目标函数和所述约束条件求解出干扰优化策略。
  13. 根据权利要求12所述的干扰优化方法，其中，利用构建的深度强化学习算法，根据所述目标函数和所述约束条件求解出干扰优化策略，包括：
    配置动作集与状态集；
    构建策略网络和动作评估网络，所述策略网络和动作评估网络采用相同的三层网络结构；
    利用所述策略网络，根据所述动作集与状态集，基于policy策略确定策略；
    利用所述动作评估网络，对所述策略进行收敛求解，得出所述干扰优化策略。
  14. 根据权利要求13所述的干扰优化方法，其中，所述动作集包括所述施扰站调整干扰的波束SSB方向、所述施扰站降低发射功率3dB、所述施扰站降低发射功率6dB、关闭所述施扰站产生干扰的时隙资源和关闭所述受扰站产生干扰的时隙资源，所述状态集包括所述受扰站上报的性能数据和MR数据中所调度的PRB调度资源、RSRP和上行吞吐量。
  15. 根据权利要求1所述的干扰优化方法，其中，在将所述干扰优化策略下发至所述施扰站和/或所述受扰站之后，所述干扰优化方法还包括：
    采集所述受扰站和所述受扰小区的优化后指标数据；
    调用所述MDA功能，基于所述优化后指标数据判断是否存在干扰，以实现优化反馈。
  16. 一种干扰优化装置，所述干扰优化装置设置在网管系统侧，包括：
    数据采集模块，用于采集指标数据；
    判断受扰模块，用于调用管理数据分析MDA功能构造的聚类算法，基于所述指标数据判断是否存在干扰；
    确定受扰站模块，用于在存在干扰的情况下确定出受扰站和受扰小区；
    确定施扰站模块，用于在存在干扰的情况下，调用所述MDA功能确定宏基站，根据所述受扰小区和所述宏基站确定施扰站；
    生成策略模块，用于调用所述MDA功能构造的强化学习算法，基于所述受扰站的受扰站信息和所述施扰站的施扰站信息生成干扰优化策略；
    下发模块，用于将所述干扰优化策略下发至所述施扰站或所述受扰站中的至少一 个，以实现干扰优化。
  17. 一种非易失性计算机可读存储介质，其上存储有计算机程序，所述程序被处理器执行时实现如权利要求1至15任一项所述的干扰优化方法。
  18. 一种电子设备，包括：
    一个或多个处理器；
    存储装置，用于存储一个或多个程序，当所述一个或多个程序被所述一个或多个处理器执行时，使得所述一个或多个处理器实现如权利要求1至15任一项所述的干扰优化方法。
  19. 一种计算机程序，包括：
    指令，所述指令当由处理器执行时使所述处理器执行权利要求1至15任一项所述的干扰优化方法。
  20. 一种计算机程序产品，包括指令，所述指令当由处理器执行时使所述处理器执行权利要求1至15任一项所述的干扰优化方法。
PCT/CN2022/103772 2021-09-08 2022-07-05 干扰优化方法及装置、存储介质及电子设备 WO2023035750A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111050926.8 2021-09-08
CN202111050926.8A CN115802398A (zh) 2021-09-08 2021-09-08 干扰优化方法及装置、存储介质及电子设备

Publications (1)

Publication Number Publication Date
WO2023035750A1 true WO2023035750A1 (zh) 2023-03-16

Family

ID=85416805

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/103772 WO2023035750A1 (zh) 2021-09-08 2022-07-05 干扰优化方法及装置、存储介质及电子设备

Country Status (2)

Country Link
CN (1) CN115802398A (zh)
WO (1) WO2023035750A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117349037A (zh) * 2023-12-06 2024-01-05 中国科学院深圳先进技术研究院 在离线应用干扰消除方法、装置、计算机设备及存储介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115996409B (zh) * 2023-03-23 2023-07-04 中国移动通信有限公司研究院 远端干扰检测处理方法、装置、设备及可读存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108111248A (zh) * 2017-12-29 2018-06-01 广州市贝讯通信技术有限公司 一种大气波导干扰优化方法和系统
CN110913331A (zh) * 2019-11-08 2020-03-24 中睿通信规划设计有限公司 一种基站干扰源定位系统和方法
CN112543411A (zh) * 2019-09-20 2021-03-23 中兴通讯股份有限公司 一种无线通信系统的干扰定位方法、装置和系统
CN112818596A (zh) * 2021-01-29 2021-05-18 广东振子电子科技有限公司 一种智能干扰识别定位方法及其系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108111248A (zh) * 2017-12-29 2018-06-01 广州市贝讯通信技术有限公司 一种大气波导干扰优化方法和系统
CN112543411A (zh) * 2019-09-20 2021-03-23 中兴通讯股份有限公司 一种无线通信系统的干扰定位方法、装置和系统
CN110913331A (zh) * 2019-11-08 2020-03-24 中睿通信规划设计有限公司 一种基站干扰源定位系统和方法
CN112818596A (zh) * 2021-01-29 2021-05-18 广东振子电子科技有限公司 一种智能干扰识别定位方法及其系统

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117349037A (zh) * 2023-12-06 2024-01-05 中国科学院深圳先进技术研究院 在离线应用干扰消除方法、装置、计算机设备及存储介质
CN117349037B (zh) * 2023-12-06 2024-04-19 中国科学院深圳先进技术研究院 在离线应用干扰消除方法、装置、计算机设备及存储介质

Also Published As

Publication number Publication date
CN115802398A (zh) 2023-03-14

Similar Documents

Publication Publication Date Title
WO2023035750A1 (zh) 干扰优化方法及装置、存储介质及电子设备
Keshavarznejad et al. Delay-aware optimization of energy consumption for task offloading in fog environments using metaheuristic algorithms
WO2020192306A1 (zh) 无线路由器部署方法、装置、存储介质及电子设备
CN110401964A (zh) 一种面向用户为中心网络基于深度学习的功率控制方法
CN105379412B (zh) 一种控制多个无线接入节点的系统和方法
US11928583B2 (en) Adaptation of deep learning models to resource constrained edge devices
CN113194489B (zh) 无线边缘网络中有效联邦学习的最小-最大代价优化方法
Cui et al. Optimal rate adaption in federated learning with compressed communications
WO2014101441A1 (zh) 认知协作中继转发方法及系统
Chen et al. Hybrid fog/cloud computing resource allocation: Joint consideration of limited communication resources and user credibility
Ling et al. Distributionally robust chance-constrained backscatter communication-assisted computation offloading in WBANs
Wang et al. Joint service caching, resource allocation and computation offloading in three-tier cooperative mobile edge computing system
Dou et al. Mobile edge computing based task offloading and resource allocation in smart grid
Miao et al. Genetic algorithm-based hybrid spectrum handoff strategy in cognitive radio-based internet of things
AU2020103373A4 (en) Machine learning based network intelligentization for automatically- configurable cellular communication systems
Duran et al. Asymptotic optimal control of Markov-modulated restless bandits
Sun et al. Joint ddpg and unsupervised learning for channel allocation and power control in centralized wireless cellular networks
Krishna Priya et al. Crossover-based improved sine cosine algorithm for multimedia content distribution in cloud environment
KR20230032904A (ko) 군집 좌표 매핑 기반 정의서 분류를 이용한 스마트 컨트랙트 레포지토리 구축 방법 및 시스템
Yan et al. Novel bat algorithms for scheduling independent tasks in collaborative Internet-of-Things
CN114205238A (zh) 网络资源优化、模型训练方法、装置、存储介质及设备
Liu et al. Computation offloading optimization in mobile edge computing based on HIBSA
Valerio et al. Optimising cost vs accuracy of decentralised analytics in Fog computing environments
Tong et al. D2OP: A fair dual-objective weighted scheduling scheme in Internet of Everything
Chen et al. Over-the-Air Computing Aided Federated Learning and Analytics via Belief Propagation Based Stochastic Bandits

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22866225

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE