Digital twinning of assembled telecommunications networks
By selecting simulation models and component counts based on performance indicators and dynamically assembling digital twins, the problem of insufficient speed and accuracy in mobile and telecommunications networks in the prior art is solved, and efficient simulation is achieved that is suitable for all situations.
Patent Information
- Application Number
- CN202380026835.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-03-14
- Filing Date
- 2023-03-09
- Publication Date
- 2024-11-08
AI Technical Summary
Existing digital twin technologies are difficult to find a solution for all situations in mobile and telecommunications networks, resulting in simulations being too slow or inaccurate in some cases to meet the needs of different use cases.
Dynamically assemble the digital twin to ensure that the right tradeoff between evaluation accuracy and duration is found between evaluation accuracy and duration by selecting the appropriate simulation model and component count based on the minimum evaluation accuracy and maximum evaluation duration of performance metrics in the digital twin of the telecom network. , thereby creating digital twins that are suitable for a particular situation.
It realizes a digital twin system that can operate effectively in different situations, improves simulation accuracy and efficiency, and can select the optimal number of components and simulation model complexity for specific problems, meeting the needs of various use cases.
Smart Images

Figure CN118923093A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a system for performing simulation of a digital twin of a telecommunications network.
[0002] The invention further relates to a method of performing simulation on a digital twin of a telecommunication network.
[0003] The invention also relates to a computer program product enabling a computer system to perform such a method. Background Art
[0004] The concepts of 'cyber-physical convergence' and 'digital twins' are applied / developed in various fields, including manufacturing, logistics and smart cities. The basic idea is to have a physical object with a digital representation that enables scenario-based experiments to be conducted in a manual or automated manner, for example, in order to predict and analyze the behavior / response of the object before applying the chosen adjustments to the physical object itself.
[0005] In the context of mobile networking there are various models / tools for the purposes of network planning, optimization, and technology / functional development. For example, widely used are radio network planning tools, system-level and link-level simulators, which have different degrees of modeling accuracy / complexity and evaluation speed and therefore have different scopes of use. At one extreme, link-level simulators model transmission, channel, and reception aspects in great detail and can therefore only handle a single radio link in isolation. At the other extreme, radio network planning tools typically involve fairly simplified modeling of network technologies to allow evaluation of relatively large portions of an operational network with many active users. Such tools are typically used by human experts to conduct experiments in order to derive actions that are then applied to the operational network in a non-automated manner.
[0006] In the context of self-organizing networks (SON), fully automated closed-loop control mechanisms have been developed and implemented, which are characterized by a continuous cycle of measurements, algorithmic processing of them, and application of the resulting configuration changes to the operational network. The underlying algorithms can be engineering rules of the if-then-else type, or still relatively simple artificial intelligence (AI) / machine learning (ML) based solutions, which are far less evolved and mature than the foreseeable digital twin-based solutions.
[0007] The following paper provides an overview of the concept of digital twins in the context of mobile communication networks and how they can be used for 5G network optimization: HX Nguyen, R. Trestian, D. To, and M. Tatipamula, “Digital twin for 5G and beyond,” IEEE Communications Magazine, Vol. 59, No. 2, 2021. The authors present a digital twin as consisting of a virtual model that self-learns in real time (using AI / ML techniques). The paper further describes a digital twin as an integration of different models representing different parts of a network.
[0008] Ericsson's online blog, titled "The future of digital twins: what will they mean for mobile networks?" ( https: / / www.ericsson.com / en / blog / 2021 / 7 / future-digital-twins-in-mobil e-networks ) discusses the trade-offs between complexity, accuracy, and the computational resources required to perform simulations when applying the concept of digital twins in mobile communication networks. The blog suggests using a mix of different models with varying levels of complexity depending on the use case considered, the network domain of interest, and the level of detail required. However, the blog does not disclose how this is implemented and it is not clear from the blog.
[0009] Currently, there is no “one-size-fits-all” solution for digital twinning a mobile network or other telecommunications network; the simulation performed by the digital twin may be too slow in the first case but not in the second, or too inaccurate in the first case but not in the second. Summary of the invention
[0010] A first objective of the invention is to provide a system that allows the creation of a digital twin adapted to the situation under consideration.
[0011] A second object of the invention is to provide a method that can be used to create a digital twin adapted to the situation under consideration.
[0012] In a first aspect of the present invention, a system for performing simulation on a digital twin of a telecommunications network comprises at least one processor: the at least one processor is configured to: determine a minimum evaluation accuracy and / or determine a maximum evaluation duration with respect to one or more performance indicators; select a simulation model for each of a plurality of components of the digital twin based on at least one of the minimum evaluation accuracy and the maximum evaluation duration; assemble the selected simulation models into the digital twin; and perform simulation on the digital twin.
[0013] With the system, digital twins can be assembled from more complex components and / or more components when evaluation accuracy is important and / or evaluation duration is unimportant / less important, and from less complex components and / or fewer components when evaluation duration is important and / or evaluation accuracy is less important. In this way, a digital twin suitable for the situation under consideration can be created. Examples of performance indicators are throughput, latency, and coverage probability. Simulations are typically performed using different settings of configuration parameters. Simulation models are typically determined jointly, with each component rather than (individually). Each component has a component type. For example, a telecommunications network can be a mobile communication network or a cable TV network.
[0014] The at least one processor may be configured to select a simulation model for each of the plurality of components of the digital twin based on at least one of a minimum evaluation accuracy and a maximum evaluation duration by determining a number of the plurality of components based on at least one of the minimum evaluation accuracy and the maximum evaluation duration, and selecting a simulation model for each of the plurality of components. Thus, in cases where evaluation accuracy is important and / or evaluation duration is unimportant / less important, the digital twin may be assembled from more components. For example, a relatively large number of components may be selected for component types that are most relevant to the problem under consideration.
[0015] Optionally, for at least one of the multiple components, the component type of the corresponding component can be associated with multiple simulation models of different complexity. Then, the at least one processor can be configured to select a simulation model for each of the multiple components of the digital twin based on at least one of the minimum evaluation accuracy and the maximum evaluation duration in the following manner: based on at least one of the minimum evaluation accuracy and the maximum evaluation duration, at least one simulation model is selected (jointly) from the multiple simulation models for the at least one component. Therefore, in cases where evaluation accuracy is important and / or evaluation duration is not important / less important, a digital twin can be assembled from more complex components.
[0016] The at least one processor may be configured to perform simulations on the digital twin to determine one or more network configuration parameter settings of the telecommunications network, and to configure the telecommunications network using the one or more network configuration parameter settings. This enables configuration of the operational telecommunications network with better network configuration parameter settings. These simulations may start from the current live network configuration, or where unavailable, may start from, for example, default parameter settings, vendor recommended parameter settings, intermediate range parameter settings, or arbitrary parameter settings. For example, the operational telecommunications network may be a test network or a commercial network.
[0017] For example, the one or more network configuration parameter settings may include antenna tilt settings and / or switching threshold settings and / or transmit power settings. For example, the digital twin of the telecommunications network may additionally or alternatively be used to evaluate the addition of a new site at a certain time and location. In other words, it may be learned from the simulation that the new site is beneficial.
[0018] In the case of a mobile communications network, the multiple components of this digital twin of the telecommunications network may represent network elements in the radio, core and / or cloud domains (e.g. one or more base stations and / or one or more antennas), as well as aspects of user behavior (e.g. user mobility), device / traffic characteristics and / or the propagation environment.
[0019] The at least one processor may be configured to: select a problem definition from a plurality of problem definitions, each of the plurality of problem definitions specifying one or more configuration parameters to be optimized and one or more corresponding performance indicators; further select a simulation model for each of a plurality of components of the digital twin based on the selected problem definition; and select one or more configuration parameters for simulation according to the selected problem definition. The problem definition may further specify a geographic scope and / or a technical scope. The selected problem definition may include a minimum evaluation accuracy and / or a maximum evaluation duration.
[0020] In this way, the optimal number of components of a digital twin and / or the optimal complexity of the simulation models of the components can be selected for a particular problem. This solves the problem of conventional "one size fits all" solutions, where a given digital twin may be modeled in too much detail, thus producing unacceptably slow simulations for one problem, or not modeled in enough detail, thus failing to produce sufficiently accurate simulations for another problem.
[0021] The defined problem can be an optimization problem or an evaluation problem. For example, the goal of an optimization problem can be to optimize a configuration parameter (e.g., tilt setting) in a mobile communication network, while the goal of an evaluation problem can be to evaluate the performance of a vendor-implemented feature in a mobile communication network under different settings of a certain configuration parameter. Feature evaluation inherently also involves the optimization of configuration parameters in order to quantify the benefit of the feature (i.e., the maximum achievable performance gain assuming the optimized configuration).
[0022] For example, in an optimization problem, the problem may be to optimize a configuration parameter for a single performance indicator or for multiple performance indicators (e.g., a weighted average). The performance indicator may be a combination of multiple performance indicators. The optimization of the configuration parameters may be performed under certain constraints, such as under the same constraints used to select the simulation model. For example, when selecting a simulation model for each of the multiple components based on a minimum evaluation accuracy, the optimization algorithm may use the same minimum evaluation accuracy as a constraint.
[0023] The at least one processor may be configured to: assemble the simulation model into different digital twin candidates; perform multiple simulations on each of the assembled different digital twin candidates to determine training samples; generate utility functions about one or more performance indicators for multiple digital twin candidates based on the training samples, the multiple digital twin candidates including the assembled different digital twin candidates; generate a cost function for the multiple digital twin candidates based on the training samples; and select a simulation model for each of the multiple components of the digital twin by applying the utility function to the minimum evaluation accuracy and / or by applying the cost function to the maximum evaluation duration. This is an advantageous method for selecting a simulation model for each of the multiple components of the digital twin based on at least one of the minimum evaluation accuracy and the maximum evaluation duration;
[0024] The utility function is also called the accuracy function. The cost function reflects the evaluation effort, such as the number of evaluations required to determine the best / optimized settings for the configuration parameters multiplied by the computational time required to evaluate a single configuration of the digital twin. For the system performing the simulation, the computational time required to evaluate a single configuration is determined / learned. For example, the utility function and cost function can be generated based on training samples using AI / ML techniques.
[0025] In the case of, for example, an optimization problem, the optimization problem can be solved for each of a plurality of initial settings of the configuration parameters for each digital twin candidate, e.g., determining how many simulations / iterations need to be performed for each initial configuration in order to reach the optimal / optimized configuration. These initial configuration parameter settings can be selected based on expert knowledge, for example. By using AI / ML techniques, utility and cost functions can be generated without data on every possible (initial) setting of the configuration parameters.
[0026] It may not be feasible to perform multiple simulations for each digital twin candidate that can be assembled. Therefore, digital twin candidates (from which training samples are obtained) can be selected from a reduced set of all possible digital twins that can be assembled based on expert knowledge. This expert knowledge can be related to the selected problem (domain) and / or the AI / ML technology used.
[0027] Expert knowledge may also be used in a later stage. The at least one processor may be configured to select a simulation model for each of the plurality of components by selecting a digital twin candidate from a plurality of digital twin candidates based on at least one of a (required) minimum evaluation accuracy and a maximum (acceptable) evaluation duration and further based on expert knowledge. This may help select the optimal digital twin candidate for the problem under consideration taking into account constraints.
[0028] In other words, a digital twin can be selected from a reduced set of all possible digital twins that can be assembled based on expert knowledge. The expert knowledge can be related to the selected problem (domain) and can be the same as the expert knowledge used to select the digital twin candidates, or it can be (partially) different. Selecting a simulation model for each component of the digital twin can include selecting one of the digital twin candidates for which training samples have been obtained, but this is not required. For example, utility function values and cost function values can also be determined for digital twin candidates that have not yet been simulated, for example by using AI / ML techniques.
[0029] The training sample may include the value of the one or more performance indicators, and the at least one processor may be configured to generate a cost function and / or a utility function based on the value of the one or more performance indicators. For example, the value of the performance indicator obtained for a certain digital twin candidate may be compared with the value of the performance indicator obtained for the most accurate digital twin candidate to generate a utility function.
[0030] The at least one processor may be configured to select a simulation model for each of the plurality of components of the digital twin based on a maximum (acceptable) evaluation duration, such that the simulations of the digital twin have a maximized evaluation accuracy without exceeding the maximum (acceptable) evaluation duration. This is beneficial if it is sufficient not to exceed the maximum (acceptable) evaluation duration, and it is not necessary to obtain the shortest evaluation duration; more importantly, the highest evaluation accuracy is obtained under the constraint of the evaluation duration. This is a trade-off between evaluation duration and evaluation accuracy.
[0031] Alternatively, the at least one processor may be configured to select a simulation model for each of the plurality of components of the digital twin based on a (required) minimum evaluation accuracy, such that the simulations of the digital twin have a minimized evaluation duration while satisfying the (required) minimum evaluation accuracy. This is beneficial if satisfying the (required) minimum evaluation accuracy is sufficient, and it is not necessary to obtain the highest evaluation accuracy; more importantly, a short evaluation duration is obtained under the constraint of evaluation accuracy. This is another trade-off between evaluation duration and evaluation accuracy.
[0032] In a second aspect of the present invention, a system for generating utility functions and cost functions for digital twin candidates of a telecommunications network includes at least one processor configured to: assemble a simulation model into different digital twin candidates of the telecommunications network; perform multiple simulations on each of the different assembled digital twin candidates to determine training samples; generate utility functions about one or more performance indicators for multiple digital twin candidates based on these training samples, and the multiple digital twin candidates include the different assembled digital twin candidates; generate cost functions for the multiple digital twin candidates based on these training samples; and store information specifying the cost functions and utility functions generated for the multiple digital twin candidates in a memory.
[0033] In a third aspect of the invention, a computer-implemented method for performing simulation on a digital twin of a telecommunications network comprises: determining a minimum evaluation accuracy and / or determining a maximum evaluation duration for one or more performance indicators; selecting a simulation model for each of a plurality of components of the digital twin based on at least one of the minimum evaluation accuracy and the maximum evaluation duration; assembling the selected simulation models into the digital twin; and performing simulation on the digital twin. The method may be performed by software running on a programmable device. This software may be provided as a computer program product.
[0034] In a fourth aspect of the invention, a computer-implemented method for generating a utility function and a cost function for a digital twin candidate of a telecommunications network comprises: assembling a simulation model into different digital twin candidates of the telecommunications network; performing multiple simulations on each of the assembled different digital twin candidates to determine training samples; generating utility functions for one or more performance indicators for multiple digital twin candidates based on these training samples, the multiple digital twin candidates including the assembled different digital twin candidates; generating a cost function for the multiple digital twin candidates based on these training samples; and storing information specifying the cost function and the utility function generated for the multiple digital twin candidates in a memory. The method may be performed by software running on a programmable device. This software may be provided as a computer program product.
[0035] In addition, a computer program for executing the method described herein and a non-transitory computer-readable storage medium storing the computer program are provided.For example, the computer program can be downloaded from or uploaded to an existing device, or stored when manufacturing these systems.
[0036] A non-transitory computer-readable storage medium stores at least a first software code portion that, when executed or processed by a computer, is configured to perform executable operations for performing simulations on a digital twin of a telecommunications network.
[0037] These executable operations include: determining a minimum evaluation accuracy and / or determining a maximum evaluation duration with respect to one or more performance indicators; selecting a simulation model for each of multiple components of the digital twin based on at least one of the minimum evaluation accuracy and the maximum evaluation duration; assembling the selected simulation models into the digital twin; and performing simulation on the digital twin.
[0038] A non-transitory computer-readable storage medium stores at least a second software code portion that, when executed or processed by a computer, is configured to perform executable operations for generating a utility function and a cost function for a digital twin candidate of a telecommunications network.
[0039] These executable operations include: assembling a simulation model into different digital twin candidates of a telecommunications network; performing multiple simulations on each of the assembled different digital twin candidates to determine training samples; generating utility functions regarding one or more performance indicators for multiple digital twin candidates based on these training samples, the multiple digital twin candidates including the assembled different digital twin candidates; generating a cost function for the multiple digital twin candidates based on these training samples; and storing in a memory information specifying the cost function and utility function generated for the multiple digital twin candidates.
[0040] As will be appreciated by those skilled in the art, aspects of the present invention may be implemented as devices, methods or computer program products. Therefore, aspects of the present invention may take the form of complete hardware embodiments, complete software embodiments (including firmware, resident software, microcode, etc.) or embodiments of combined software and hardware aspects that may generally all be referred to as "circuit", "module" or "system" herein. The functions described in this disclosure may be implemented as algorithms executed by a processor / microprocessor of a computer. In addition, aspects of the present invention may also take the form of a computer program product implemented in one or more computer-readable media, the computer-readable medium having a computer-readable program code embodied thereon (e.g., stored).
[0041] Any combination of one or more computer-readable media can be used.Computer-readable media can be computer-readable signal media or computer-readable storage media.Computer-readable storage media can be, for example, but not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing.More specific examples of computer-readable storage media can include, but are not limited to, the following: electrical connections with one or more wires, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.In the context of the present invention, computer-readable storage media can be any tangible medium that can contain or store programs used by or in combination with instruction execution systems, devices, or equipment.
[0042] A computer-readable signal medium may include a propagated data signal having a computer-readable program code embodied therein (e.g., in baseband or as part of a carrier wave). Such a propagated signal may take any of a variety of forms, including but not limited to electromagnetic, optical, or any suitable combination thereof. A computer-readable signal medium may be any computer-readable medium that is not a computer-readable storage medium and that can communicate, propagate, or convey a program (for use by or in connection with an instruction execution system, device, or apparatus).
[0043] Can use any suitable medium transmission to be embodied in the program code on computer readable medium, this medium includes but is not limited to any suitable combination of wireless, wired, optical fiber, cable, RF etc. or aforementioned.The computer program code for performing the operation of each aspect of the present invention can be written in any combination of one or more programming languages, and these one or more programming languages include such as Java (TM), Smalltalk, C++ and such as conventional process programming languages such as " C " programming language or similar programming languages.Program code can be performed on the user's computer completely, also can be partly performed on the user's computer (as independent software package), can also be partly performed on the user's computer and partly performed on the remote computer, or can be performed completely on the remote computer or server.In the latter scenario, the remote computer can be connected to the user's computer through any type of network, and this network includes local area network (LAN) or wide area network (WAN), or can be connected to an external computer (for example, by using the Internet of Internet Service Provider).
[0044] Aspects of the present invention are described below with reference to the flowchart and / or block diagram of the method, device (system) and computer program product according to an embodiment of the present invention. It should be understood that each frame of the flowchart and / or block diagram and the combination of frames in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device, specifically a microprocessor or a central processing unit (CPU), to produce a machine so that instructions executed via a processor of a computer, other programmable data processing device or other equipment create a device for implementing the function / action specified in one or more frames of the flowchart and / or block diagram.
[0045] These computer program instructions may also be stored in a computer-readable medium, which can instruct a computer, other programmable data processing device, or other equipment to function in a particular manner, so that the instructions stored in the computer-readable medium produce a manufactured article including instructions, which implement the functions / actions specified in one or more boxes of the flowchart and / or block diagram.
[0046] The computer program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus, or other device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable apparatus provide a process for implementing the functions / actions specified in the flowchart and / or one or more block diagram blocks.
[0047] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functions and operations of possible implementations of devices, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram may represent a module, a fragment or a code portion, which includes one or more executable instructions for implementing (multiple) specific logical functions.
[0048] It should also be noted that in some alternative embodiments, the functions indicated in the blocks may not occur in the order indicated in the figure. For example, depending on the functions involved, two blocks shown in succession may actually be executed substantially simultaneously, or sometimes the two blocks may be executed in reverse order. It should also be noted that each block of the block diagram and / or flow chart and the combination of blocks in the block diagram and / or flow chart may be implemented by a dedicated hardware-based system or a combination of dedicated hardware and computer instructions that performs specific functions or actions. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] These and other aspects of the invention will become apparent and will be further elucidated by way of example with reference to the accompanying drawings, in which:
[0050] Figure 1 is a flow chart of a first embodiment of a method for performing simulation;
[0051] Figure 2 is a flow chart of a second embodiment of a method for performing simulation;
[0052] Figure 3 is a flow chart of a third embodiment of a method for performing simulation;
[0053] Figure 4 Shown in Figure 3 Digital twins assembled in the method;
[0054] Figure 5 The different phases that can be distinguished in terms of the steps of the method are shown;
[0055] Figure 6 A flowchart showing an embodiment of a method of generating a cost function and a utility function and a fourth embodiment of a method of performing a simulation;
[0056] Figure 7 and Figure 8 Two example uses of example cost and utility functions are shown;
[0057] Fig. 9 is a block diagram of an embodiment of a system; and
[0058] Fig.10 is a block diagram of an exemplary data processing system for executing the method of the present invention.
[0059] Corresponding elements in the drawings are denoted by the same reference numerals. DETAILED DESCRIPTION
[0060] Figure 1 A first embodiment of a computer-implemented method for performing simulation on a digital twin of a telecommunications network is shown in . Step 101 comprises at least one of the following: a) determining a (required) minimum evaluation accuracy with respect to one or more performance indicators and b) determining an (acceptable) maximum evaluation duration. Step 103 comprises selecting a simulation model for each of a plurality of components of the digital twin based on at least one of the minimum evaluation accuracy and the maximum evaluation duration as determined in step 101. Step 105 comprises assembling the simulation model selected in step 103 into the digital twin.
[0061] Step 107 includes performing a simulation on the digital twin assembled in step 105. For example, a simulation can be performed to find one or more optimal configuration parameter settings. For example, a gradient-based optimization strategy can be used. For example, the simulation model can be a black box that can set / adjust certain parameter settings.
[0062] Step 101, step 103, and step 105 are part of the assembly phase 152 of the digital twin, and step 107 is part of the utilization phase 153 of the digital twin. According to some definitions of the term "digital twin", the digital twin is updated based on real-time data. In the context of this specification, this is possible, but not required.
[0063] exist Figure 1 In the embodiment of , step 103 is implemented by step 121 and step 123. Step 121 includes determining the number of the plurality of components based on at least one of a minimum evaluation accuracy and a maximum evaluation duration. Figure 1 In an embodiment of the present invention, the number of multiple components is determined in step 121 by determining the number of components of each component type. Optionally, one or more (but not all in this embodiment) component types have a fixed number of components. The number of components of each remaining component type is determined (jointly) based on at least one of a minimum evaluation accuracy and a maximum evaluation duration. Step 121 may include summing the number of components of each component type, but this is not required.
[0064] Step 123 includes selecting a simulation model for each of the plurality of components. Figure 1 In an embodiment of the invention, each component type is associated with a single simulation model. Thus, the same simulation model is selected for each component of the same component type. As an example, a digital twin of a radio access network may use component types A) base station, B) antenna, C) propagation environment, and D) user mobility. If two A-type components, four B-type components, one C-type component, and one D-type component are determined in step 121, then eight simulation models are selected in step 123, where two A-type simulation models are the same and four B-type simulation models are the same.
[0065] Figure 2 A second embodiment of a computer-implemented method for performing simulations on a digital twin of a telecommunications network is shown in FIG. Figure 2 The method also includes step 101, step 105, and step 107, similar to Figure 1 However, in Figure 2 In the embodiment, step 103 is implemented by steps 131 and 133, and for at least one of the plurality of components, the component type of the corresponding component is associated with a plurality of simulation models of different complexity.
[0066] Step 131 includes determining the number of the plurality of components. Figure 2In an embodiment of the invention, a fixed number of components is determined for all component types, i.e. a number of components that does not depend on the minimum evaluation accuracy or the maximum evaluation duration. The number of components may still depend on the selected problem definition, such as Figure 6 Described. Figure 2 In a variation of the embodiment, step 131 is omitted.
[0067] Step 133 includes selecting a simulation model for each of the components having a component type associated with multiple simulation models (jointly) based on the minimum evaluation accuracy and / or the maximum evaluation duration. Step 133 further includes selecting a simulation model for each of the components having a component type associated with only one simulation model (if such a component type exists). If the same simulation model is selected for each component of the same component type, step 131 may alternatively be performed after step 133 or in parallel with step 133.
[0068] Figure 3 A third embodiment of a computer-implemented method for performing simulations on a digital twin of a telecommunications network is shown in FIG. Figure 3 The method also includes step 101, step 105, and step 107, similar to Figure 1 and Figure 2 However, in Figure 3 In the embodiment of the present invention, step 103 is performed by Figure 1 Step 121 and Figure 2 Step 133 is implemented.
[0069] Figure 4 Shown in Figure 3 Digital twins assembled in a method. Figure 4 Four component types 11, 21, 31, and 41 are shown. Figure 4 In the example shown in Figure 2, each component type is associated with three different simulation models with different complexity (low, medium, and high). These simulation models are stored in a simulation model repository for the digital twin component type, which has well-defined inter-component interfaces to support custom assembly. Figure 4 The digital twin of is just one example: for example, there may be more or less than four component types, and / or more or less than three levels of complexity.
[0070] Component type 11 is associated with low complexity simulation model 13, medium complexity simulation model 14, and high complexity simulation model 15. Component type 21 is associated with low complexity simulation model 23, medium complexity simulation model 24, and high complexity simulation model 25. Component type 31 is associated with low complexity simulation model 33, medium complexity simulation model 34, and high complexity simulation model 35. Component type 41 is associated with low complexity simulation model 43, medium complexity simulation model 44, and high complexity simulation model 45.
[0071] exist Figure 4 In the example, Figure 3 In step 121, two components 3 and 4 are determined for component type 11, two components 5 and 6 are determined for component type 21, one component 7 is determined for component type 31, and one component 8 is determined for component type 41. These six components 3-8 are Figure 4 Shown in , as a component of the digital twin 1.
[0072] exist Figure 4 In the example, Figure 3 In step 133, simulation models of different complexity are selected for two components 3 and 4 of component type 11, namely, low complexity simulation model 13 and medium complexity simulation model 14, while the same simulation model, namely, low complexity simulation model 23, is selected for two components 5 and 6 of component type 21. In addition, low complexity simulation model 33 is selected for single component 7 of component type 31, and low complexity simulation model 43 is selected for single component 8 of component type 41. These simulation models are Figure 4 Shown in the figure as assembled in the digital twin 1.
[0073] exist Figures 1 to 3 In an embodiment, the simulation model is selected based only on the minimum evaluation accuracy and / or the maximum evaluation duration. In an alternative embodiment, the simulation model can be selected based on expert knowledge of the problem that has arisen, such as Figure 6 For example, the expert knowledge may be used to determine a preferred minimum and / or maximum number of components for each component type and / or a preferred minimum and / or maximum complexity of the simulation model. For example, the expert knowledge may specify that for a given problem, a user mobility model of high complexity is required and / or a minimum of 57 base stations is required.
[0074] Illustrative examples of digital twin components with varying levels of complexity include:
[0075] • Base Station - At one extreme, a fully functional replica of a base station may be used (perhaps provided by the equipment vendor), while at the other extreme the base station may be reduced to just a combination of a few key radio resource management mechanisms.
[0076] Antennas – If the problem being solved involves optimization of beamforming parameters, a detailed model of the deployed antenna array may be required, including, among other things, the number of antenna elements, subarray structure, and element-level antenna diagrams, while for other problems it may be sufficient to reduce the antenna array to just a single active transmit / receive antenna and appropriately configured antenna diagrams.
[0077] • Propagation environment - For problems involving beamforming or channel adaptive packet scheduling, it is important to fully incorporate multipath fading into the propagation model, while for other problems it may be sufficient to exclude multipath fading completely and model only path loss and possibly shadow fading.
[0078] User mobility - When optimizing handover or dynamic beam steering parameters, it may be important to consider a detailed user-level mobility model. In other problems, it may be sufficient to have a macroscopic model of, for example, traffic flow between industrial and residential areas, while in yet other problems, user mobility may be completely ignored.
[0079] Other examples include user equipment, spatiotemporal traffic characteristics, and various core network functions and interfaces.
[0080] As about Figure 1 As described, Figure 1 An assembly phase 152 of the digital twin and an utilization phase 153 of the digital twin are shown. Figure 5 A preparation phase 151 is further shown preceding the assembly phase 152. The assembly phase 152 and the utilization phase 153 are online phases, whereas the preparation phase 151 is typically an offline phase. The preparation phase 151 covers the development of simulation models of the type of components of the digital twin of the telecommunications network in question, possibly involving measurements made on the telecommunications network and AI / ML methods in the modeling itself. In an embodiment, the preparation phase 151 is also responsible for generating various utility functions or cost functions (e.g. related to evaluation accuracy and evaluation duration), see e.g. Figure 6 .
[0081] Since the telecommunication network may continuously change depending on the context, the preparation phase 151 may be a continuous process of developing / adjusting simulation models of component types. The assembly phase 152 handles the customized assembly of the digital twin by selecting the relevant models built in the preparation phase 151. In the utilization phase 153, the assembled digital twin is used for simulation, for example, to evaluate the impact of candidate configurations / actions and implement the selected configurations / actions on the twin telecommunication network.
[0082] exist Figures 1 to 3 In the embodiment of the present invention, the utilizing stage 153 includes step 107. Step 107 includes performing a simulation on the digital twin assembled in step 105. Figures 1 to 3In a variation of the embodiment of , step 107 is implemented by step 141, and the utilization stage 153 further includes step 143. Figure 5 Shown in.
[0083] Step 141 includes performing a simulation on the digital twin assembled in step 105 to determine one or more network configuration parameter settings of the telecommunication network. Examples of configuration parameters of the radio access network of the mobile communication network are handover thresholds, scheduling weights, admission control thresholds, congestion control parameters, massive MIMO / beamforming parameters, CSI feedback configuration parameters, and slicing parameters. Examples of configuration parameters of the core network of the mobile communication network are slicing parameters, routing policy parameters, QoS management parameters, charging policy parameters, service area restriction parameters, PLMN selection parameters, and paging policy parameters. Step 143 includes configuring the twin telecommunication network using the one or more network configuration settings determined in step 141.
[0084] Figure 6 An embodiment of a computer-implemented method for generating a cost function and a utility function and a fourth embodiment of a computer-implemented method for performing simulation on a digital twin of a telecommunications network are shown. The method for generating a cost function and a utility function is performed in a preparation phase 151. The method for performing simulation on a digital twin, including steps 101 to 107, is performed in an assembly phase 152 and an utilization phase 153.
[0085] As mentioned previously, the preparation phase 151 covers the development of simulation models of the digital twin component types of the telecommunications network in question, possibly involving measurements made on the telecommunications network and the application of AI / ML methods in the modeling itself. This can be performed using conventional techniques. It is assumed that a repository of simulation models of the digital twin component types is available, with well-defined inter-component interfaces to support assembly.
[0086] Preparation phase 151 includes step 171, step 173, step 175, step 177, step 178, step 179, step 181, step 183, and step 187. In a first iteration of step 171, step 171 includes selecting a first problem definition from a plurality of problem definitions. Each of the plurality of problem definitions specifies one or more configuration parameters to be optimized and one or more corresponding performance indicators. Optionally, one or more performance indicators that need to achieve minimum and / or maximum values can be included in the problem definition (as constraints). Examples of performance indicators are throughput and latency.
[0087] Examples of configuration parameters are already available for Figure 5The problem definition may further specify a geographic scope and / or a technical scope. The problem definition may further specify a (required) minimum evaluation accuracy and / or a (acceptable) maximum evaluation duration, but if not, these may be entered separately during the assembly phase 152.
[0088] For example, a problem definition may specify an optimization or evaluation problem. In both cases, a simulation is performed to find one or more optimal / optimized configuration parameter settings (usually with certain constraints) with respect to a performance metric. The goal of solving an optimization problem is to find the one or more optimal / optimized configuration parameter settings. The goal of solving an evaluation problem is to find the value of the performance metric that corresponds to the one or more optimal / optimized configuration parameter settings.
[0089] In the assembling phase 152, problems may be triggered periodically as part of an automated network optimization cycle, may be triggered by accidental events such as base station failures, or may be triggered by a network engineer activating an evaluation of a vendor-implemented feature in a mobile operator's network. When a problem occurs in the assembling phase 152, a corresponding problem definition is selected from the plurality of problem definitions (e.g., from a list of problem definitions).
[0090] With respect to the target performance indicator, a well-defined optimization / evaluation problem should preferably include a clear definition of what needs to be optimized / evaluated, which may be, for example, just a single KPI (e.g., coverage probability), multiple KPIs (e.g., X ,K Y ), that is, K = α X K X +β Y K Y , (e.g. 0.8×packet loss rate+0.2×switching ping-pong ratio), or a combination of optimized KPI and conditional KPI, i.e. in refers to an indicator function (eg, optimizing average user throughput with a coverage probability greater than 99%). The term "one or more performance indicators" covers all these options.
[0091] In a first iteration of step 173, step 173 includes selecting a simulation model for each of the plurality of components of the first candidate digital twin. Step 173 may be similar to Figures 1 to 3Thus, a desired number of components (e.g., per component type) may be determined in step 173 and / or a simulation model having a desired complexity (e.g., per component or component type) may be determined in step 173. The desired number of components may depend on the complexity of the selected simulation model. In a first iteration of step 175, step 175 includes assembling the simulation models of the plurality of components into a first digital twin candidate. Step 175 may be similar to Figures 1 to 3 Step 105.
[0092] Each (candidate) digital twin can be represented by a vector n Represents. Vector n Indicates the number of components in the digital twin for each twin component type available in the repository and their associated complexity level (e.g. n = number of BSs at high complexity level, number of BSs at medium complexity level, number of BSs at low complexity level, number of antennas at high complexity level, number of antennas at medium complexity level, number of antennas at low complexity level, etc.). Some of them can be set to 0.
[0093] Step 177 includes performing multiple simulations on the digital twin candidates assembled in step 175 to determine training samples. n , iterative simulation is performed using various settings of the one or more configuration parameters (e.g., a certain range of inclination settings). The training samples usually include the values of the one or more performance indicators. n , step 177 may include the following two sub-stages:
[0094] Determine settings for the one or more configuration parameters (e.g., a range of tilt settings to evaluate for a particular problem). Expert knowledge may be used to select settings for the one or more configuration parameters, i.e., select candidate configurations. This sub-stage may include selecting an initial configuration.
[0095] ·For a given n and candidate configurations for actual simulation. nDifferent runs are performed. Each run comprises a number of iterations. (Pseudo) random values may be used in the simulation. Therefore, preferably, a sufficient number of independent replications with different random seeds are performed as part of each iteration in order to achieve the required statistical accuracy for the (multiple) performance indicators. Different sets of replications may be performed for one or more load levels / scenarios. One or more runs are performed starting from one or more different initial configurations. In each run, iterations of the optimization algorithm are performed until a stopping criterion of the optimization algorithm is met. After an iteration with the initial configuration and after each next iteration, the optimization algorithm determines which is the next configuration to be tried. After the stopping criterion of the optimization algorithm is met, the total time required to determine the optimized configuration (the number of simulations / iterations multiplied by the average time per simulation / iteration) and the corresponding estimate of the performance indicator (considered to be statistically reliable) are stored / output in relation to a given n. If multiple runs are performed starting from multiple different initial configurations, it is possible to compare the results with a given n. n The average total time of each run is stored / outputted accordingly. The individual outputs of this sub-phase can be used as training samples (e.g. [question, n , KPI name]->[KPI value, total time]) to learn the cost function and utility function. If the optimized configuration is at the edge of a range of candidate configurations, the range can be adjusted to cover candidate configurations outside the edge. In this case, the first sub-stage and the second sub-stage can be repeated, and other candidate / initial configurations can be selected in the first sub-stage. An example of an optimization algorithm that can be used is gradient-based optimization in multidimensional space. The separate simulation model for each component type is typically a black box.
[0096] Step 178 includes checking whether it is necessary to perform simulation on another digital twin candidate, i.e., whether simulation has been performed on a sufficient number of digital twin candidates. Expert knowledge is preferably used to limit the set of digital twin candidates from which digital twin candidates are selected in step 173. In this case, step 178 may include checking whether a sufficient number of digital twin candidates have been selected from the set to generate suitable cost functions and utility functions with a reasonable amount of effort.
[0097] If simulation needs to be performed on another digital twin candidate, step 173 is repeated, and the method is as follows Figure 6 In the next iteration of step 173, step 173 includes selecting a simulation model for each of the plurality of components of the next candidate digital twin. In the next iteration of step 175, step 175 includes assembling the simulation models of the plurality of components into the next digital twin candidate.
[0098] Step 179 and step 181 are performed after step 178. Step 179 includes generating a utility function about the one or more performance indicators for multiple digital twin candidates based on the training samples determined in step 177. The multiple digital twin candidates include each digital twin candidate assembled in step 175. The utility function reflects the estimated accuracy about the one or more performance indicators. Step 181 includes generating a cost function for the multiple digital twin candidates based on the training samples determined in step 177. The cost function reflects the actual experienced and optionally estimated evaluation workload. Steps 179 and 181 may involve the use of machine learning techniques, for example, for estimating evaluation duration and accuracy for candidate twins for which training samples have not yet been obtained. In alternative embodiments, only a utility function or only a cost function is generated.
[0099] Typically, the utility function is determined based on the value of one or more performance indicators included in the training sample. For example, a utility function can be generated by first determining the digital twin candidate that produces the most accurate performance results among the digital twin candidates, that is, the fully deployed digital twin. The fully deployed digital twin takes into account the most complex simulation model available for each component type and takes into account the entire network or a subset thereof that produces the same evaluation accuracy. The utility function can be determined by normalization so that the function result of the digital twin candidate is 1, while the function results of other digital twin candidates are between 0 and 1. The utility function result value of a digital twin candidate reflects the utility of the digital twin candidate relative to the digital twin candidate with the highest utility (i.e., result value 1).
[0100] The cost function can be generated by, for example, determining the average or maximum total time required to determine the optimal / optimized configuration for each digital twin candidate (e.g., the average or maximum value of multiple iterations of the optimization algorithm with different initial configurations); or determining the average or maximum number of simulations / iterations required to determine the optimized configuration for each digital twin candidate, and multiplying the average or maximum number of simulations / iterations by the computational time required to perform a single simulation / iteration (e.g., in seconds).
[0101] For each considered configuration, preferably enough replications with different random seeds are performed to obtain sufficient statistical reliability of the obtained performance results.Thus, each iteration may include one or more replications. Figure 7 and Figure 8 Examples of utility functions and cost functions are shown in . Step 183 includes storing in a memory information specifying the cost function and utility function generated for the problem definition selected in step 171 .
[0102] Step 187 includes checking whether a cost function and utility function should be generated for another problem definition. If steps 173 to 183 have been performed for all of the multiple problem definitions, then the preparation phase 151 is complete (although it may be repeated later). If steps 173 to 183 have not been performed for all of the multiple problem definitions, then step 171 is repeated and the method continues as described above. Figure 6 In the next iteration of step 171 , step 171 includes selecting a next problem definition from the plurality of problem definitions.
[0103] Machine learning can be used to determine the cost function result values and utility function result values associated with digital twin candidates for which simulation has not yet been performed. Even the training samples obtained for one problem can be used to learn the cost function and utility function for another problem. In this case, steps 179, 181, and 183 can be performed after step 187 instead of before step 187. In any case, a cost function and a utility function are generated for each problem from a plurality of potential problems (e.g., a list of potential problems).
[0104] As described above, the assembly phase 152 handles the custom assembly of the digital twin by selecting the relevant models built in the preparation phase 151. Figure 6 In an embodiment, when an actual problem arises, relevant utility functions and cost functions are selected in the assembly phase 152 to determine the digital twin.
[0105] The assembly phase 152 includes step 191, step 193, step 195, step 101, step 103 and step 105. Step 191 includes selecting a problem definition from a plurality of problem definitions. In the assembly phase 152, the selected problem definition needs to correspond to a problem that occurs in an operational telecommunications network. Therefore, if the problem definition is not selected completely automatically by the system, but is selected based on user input, the user should not select any problem definition that does not correspond to the problem that occurs. For example, the operational telecommunications network may be a test network or a commercial network. Step 193 includes retrieving from a memory the generated cost function and utility function associated with the problem definition selected in step 191.
[0106] Step 101 comprises determining a (required) minimum evaluation accuracy and / or determining an (acceptable) maximum evaluation duration with respect to the one or more performance indicators specified in the selected problem definition. Figure 6 In an embodiment of the invention, the minimum evaluation accuracy and / or the maximum evaluation duration are determined separately in step 101, rather than being specified in the problem definition. In an alternative embodiment, the problem definition includes a (required) minimum evaluation accuracy and / or a (acceptable) maximum evaluation duration.
[0107] After step 193 and step 101 have been performed, step 103 is performed. Step 103 is implemented by step 195. Step 195 includes selecting a simulation model for each of a plurality of components of the digital twin by performing at least one of the following: a) applying a utility function corresponding to the problem definition retrieved in step 193 to the minimum evaluation accuracy determined in step 101, and b) applying a cost function corresponding to the problem definition retrieved in step 193 to the maximum evaluation duration determined in step 101. Step 195 may be viewed as including a step of selecting a digital twin from a plurality of digital twin candidates based on at least one of the minimum evaluation accuracy and the maximum evaluation duration, and further based on the selected problem definition.
[0108] Digital twin candidates can differ from each other by having a different number of components and / or having simulation models of different complexity, e.g. Figures 1 to 3 Described. Step 195 may include a first substep, in which a subset of the multiple digital twin candidates is selected, and a second substep, in which a digital twin is selected from the subset. For example, the subset can be selected based on expert knowledge. For example, for the tilt optimization problem, only low-complexity BSs can be considered. For example, information that will help select the subset (and represent expert knowledge) can be included in the problem definition.
[0109] The multiple digital twin candidates may include all possible digital twin candidates, and may even be infinite. In this case, the multiple digital twin candidates may include digital twin candidates whose training samples are not obtained in step 177 but whose utility values and cost values can be derived / learned from the utility values and cost values determined for other digital twin candidates. Alternatively, the multiple digital twin candidates may only include digital twin candidates for which training samples were obtained in step 177.
[0110] As mentioned above, selecting the optimal complexity and / or the optimal number of components per component type is very important for customizing digital twins, and this selection can be made based on the minimum evaluation accuracy and / or the maximum evaluation duration.
[0111] The actual evaluation duration typically depends on the available computing resources. Since higher complexity requires more computing resources (e.g., in terms of processing power or memory resources), the availability of these resources can also affect the feasibility of solving a given optimization problem within a given deadline and at a given twin complexity. Therefore, limitations on computing resources may impose restrictions on the allowed complexity. It can be assumed that the utilization phase 153 is performed using the same computing resources as the preparation phase 151, so the cost function can be used without adjustment, or it can be assumed that it is known how to convert time into actual computing resources.
[0112] The (acceptable) maximum evaluation duration typically depends on the urgency of the optimization problem. For example, if a site fails and surrounding sites need to reconfigure their antenna tilt or power settings, then finding a suboptimal solution is OK as long as it is fast enough (rapidly and roughly), thus allowing for a lower twin complexity. Alternatively, a process that periodically updates antenna tilt and power settings in response to spatiotemporal variations in the offered traffic load does not require an urgent solution, so the process can exploit a more complex twin to achieve a (near) optimal solution.
[0113] It may be necessary to make a trade-off between the number of components selected for each component type and the complexity of the components selected for each component type. This trade-off typically depends on the scope of the optimization problem and can be made with the help of expert knowledge. Some optimization problems are inherently geographically local in nature, such as the cell-specific configuration of packet scheduling parameters. Therefore, twin-based experiments can be performed for single-cell scenarios to derive optimal parameter settings, which allows the use of very detailed twins and can still be completed in a reasonable time. Alternatively, the above-mentioned antenna tilt optimization problem inherently requires a wider (regional) scope, as the tilt setting affects the inter-cell interference and jointly determines the coverage. Therefore, the twin complexity should be lower compared to the previous (local) optimization example to allow the evaluation of a large range of tilt settings for each antenna in a reasonable time.
[0114] In addition to limited geographic scope, the problem can also be characterized by limited technical / network scope, such as involving only the RAN or a more complex RAN / CN scope, or involving only a limited set of protocol layers. As with limited geographic scope, here too, the scope can affect the degree of modeling and, therefore, the degree of twin detail required.
[0115] Step 105 includes assembling the simulation models selected in step 103 (e.g., for the number of components determined in step 103 and / or the complexity determined in step 103), i.e., the simulation models associated with the digital twin selected from the plurality of digital twin candidates, to obtain a digital twin. Figure 6 In an embodiment, the assembled digital twin candidate in step 175 is not stored in the memory. In an alternative embodiment, if the digital twin selected from the plurality of digital twin candidates (e.g., in step 175) has been assembled and stored in the memory, the selected digital twin can be obtained from the memory.
[0116] Next, in the utilization phase 153, the customized digital twin assembled in the assembly phase 152 is applied to a problem-specific set of simulations to derive one or more optimal / optimized configuration parameter settings (e.g., tilt settings), which can then be applied to an operational telecommunications network, such as a test telecommunications network or a commercial telecommunications network.
[0117] Utilization phase 153 includes step 197 and step 107. Step 197 includes selecting the one or more configuration parameters specified in the problem definition selected in step 191. Step 107 includes performing simulations on the digital twin assembled in step 105 using different settings of the one or more configuration parameters selected in step 197, such as starting from the current live network configuration, or if unavailable, starting from default parameter settings, vendor recommended parameter settings, mid-range parameter settings, or arbitrary parameter settings.
[0118] The simulations performed in step 107 are similar to those performed in step 177, but only for one problem and one assembled digital twin, i.e., the selected digital twin candidate. If the (acceptable) maximum duration allows, the simulations can be performed with multiple initial configurations. For example, insights gained from a given set of simulations with one or more initial configurations may suggest new initial configurations. The optimization algorithm and stopping criteria used in step 177 can also be used in step 107. As described above, an example of an optimization algorithm that can be used is gradient-based optimization in a multidimensional space. Optionally, step 107 is implemented by step 141, which is followed by step 143, as described with respect to Figure 5 as described.
[0119] Figure 7 and Figure 8 Two example uses of example cost and utility functions are shown. While there may be situations where only a cost function or only a utility function is needed, typically selecting a digital twin from multiple digital twin candidates involves a trade-off between the accuracy of the results and the evaluation effort (i.e., the time required to determine the best / optimized configuration). Figure 6 As described, based on (e.g., periodic) offline experiments on different digital twin candidates, a relevant utility function or cost function (e.g., regarding estimated evaluation accuracy and estimated evaluation effort / duration) can be generated (e.g., machine learning).
[0120] Figure 7 and Figure 8 The utility function U is shown P ( n )61 and cost function C P ( n )62. Utility function U P( n )61, also known as the accuracy function, represents the closeness of the performance achieved by each twin option for a specific problem P relative to the performance of the fully expanded (the most realistic, closest to the real-world object and assumed to be the ground truth) twin model. Therefore, as the twin options considered become as complex as the fully expanded twin, the curve 61 converges to '1', and in fact, depending on the problem being solved, the curve can reach the '1' level for twin candidates with less complexity than the fully expanded model. The fully expanded digital twin 73 considers the most complex component models available for each component type and a sufficient number of components. Ideally, the fully expanded digital twin 73 should consider the entire network. However, in practice, a subset of the entire network may be sufficient.
[0121] Cost function C P ( n )62 is also called the evaluation workload function, which indicates the function used to determine the n The amount of evaluation effort (e.g., in seconds) required to find the optimal / optimized configuration parameter settings. Cost function C P ( n )62 can be E P ( n ) and T P ( n ). P ( n ) is the expected number of configurations that need to be simulated in order to determine the optimal configuration parameter settings (e.g., the number of tilt vectors that need to be simulated as part of the optimization algorithm in order to determine the optimal / optimized configuration parameter settings, i.e., the number of iterations; this number typically increases with the number of BSs included in the twin). P ( n ) is the n The computation time (absolute value, e.g., in seconds) required to accurately simulate a given single configuration of the digital twin, i.e., the computation time for a single iteration.
[0122] As mentioned above, the function U P ( n )、E P ( n ), and T P ( n ) depends on the problem (P). Let E P ( n) varies with the problem and can give a better estimate of the expected number of simulations required, since this will not be averaged over all potential problems (e.g. the number of possible tilt settings is different from the number of possible scheduling parameter settings, so the optimization will require a different number of simulations / iterations). For a particular problem, U can be obtained by observing or estimating (e.g. in the case of interpolation or using AI / ML models) the (possibly composite) performance metric or metrics achieved. P ( n ), and T can be obtained by calculating the simulation time required for the target (possibly composite) KPI P ( n ).
[0123] Once a periodic or event-induced problem triggers the use of a digital twin, the most appropriate twin option can be selected based on the (required) minimum evaluation accuracy and the (acceptable) maximum evaluation duration. To find the optimal digital twin, the trade-off between the accuracy of the results and the evaluation effort can be made in different ways. For example, the following options can be considered:
[0124] I. Maximize U P ( n ), the condition is C P ( n )≤C MAX ;or
[0125] II. Minimize C P ( n ), provided that U P ( n )≥U MIN .
[0126] As about Figure 6 As described in step 195, first a subset of multiple digital twin candidates can be selected based on expert knowledge, and then a digital twin can be selected from the subset. For example, the subset can be selected based on expert knowledge related to the problem. For example, for the inclination optimization problem, only low-complexity BSs are considered. The expert knowledge can be the same as or (partially) different from the expert knowledge used in the preparation stage 151. For example, in addition to knowledge related to the problem, the expert knowledge used in the preparation stage 151 can also include knowledge of machine learning.
[0127] Figure 7 Option II is shown. Figure 7 66. MINwhile having a minimized evaluation duration. As described above, the number of the plurality of components may be selected as part of selecting a simulation model for each of the plurality of components and may not be predetermined.
[0128] The digital twin candidate with the lowest evaluation duration under this condition is digital twin candidate 71, so it is selected. The maximum evaluation accuracy (i.e., 1) is represented by line 65. The fully expanded digital twin candidate 73 has the maximum evaluation accuracy. Curve 61 increases relatively quickly, and relatively more digital twin candidates have an evaluation accuracy of 1 or close to 1. Digital twin candidate 72 has almost the same evaluation accuracy as the fully expanded digital twin candidate 73. Although the fully expanded digital twin candidate 73 and the digital twin candidate 72 have at least the minimum evaluation accuracy U MIN , but they do not have a minimized evaluation duration. Digital twin candidate 70 does not meet the condition because U P ( n ) is not equal to or exceeds U MIN .
[0129] Figure 8 Option I is shown. Figure 8 67. The diagram shows the selection of a simulation model for each of the plurality of components of the digital twin based on the maximum allowed evaluation duration so that the maximum evaluation duration C represented by line 67 is not exceeded. MAX In the case of , the simulation of the digital twin will have a maximized evaluation accuracy. As described above, the number of the plurality of components may be selected as part of selecting a simulation model for each of the plurality of components and may not be predetermined.
[0130] The digital twin candidate with the highest evaluation accuracy under this condition is digital twin candidate 72, so it is selected. The fully expanded digital twin candidate 73 does not meet the condition because C P ( n ) exceeds the maximum evaluation duration of the fully deployed digital twin candidate 73. Although digital twin candidates 70 and 71 meet the condition that C P ( n ) not exceeding C MAX , but they do not maximize the evaluation accuracy.
[0131] exist Figure 7 and Figure 8In the example, only four digital twin candidates 70-73 are processed, but curves 61 and 62 represent more than four digital twin candidates. Processing digital twin candidates 70-73 does not necessarily mean that training samples have been obtained for (all) these digital twin candidates in the preparation phase. Other options besides Option I and Option II are also possible. There is no need to minimize the evaluation duration or maximize the evaluation accuracy.
[0132] Fig. 9 The present invention is a block diagram of an embodiment of a system for performing simulation on a digital twin of a telecommunications network, and an embodiment of a system for generating a utility function and a cost function for a candidate digital twin of a telecommunications network. The system for generating a utility function and a cost function includes at least device 211, and may further include device 201, for example. The system for performing simulation on a digital twin of a telecommunications network includes at least device 201, and may further include device 211, for example. System 210 includes device 201 and device 211. In Fig. 9 In the example of , the telecommunication network is a mobile communication network 200. The mobile communication network 200 comprises, among other things, a radio access network (RAN) 231 and a core network (CN) 221.
[0133] RAN 231 includes base stations 233 and 234. For example, mobile communication network 200 may be a 5G network, RAN 231 may be a 5G new radio RAN, and base stations 233 and 234 may be 5G gNodeB base stations. Each of base stations 233 and 234 may include multiple distributed units that share a common centralized unit in a centralized RAN (C-RAN) architecture. CN 221 includes core network components 223, 224, and 225.
[0134] Device 201 is used online to manage RAN 231 and CN 221, and may be part of a management network of mobile communication network 200. Device 211 is used offline to generate a utility function and a cost function. Fig. 9 In the example of FIG. 2 , mobile devices 241 and 242 are connected to base station 233 , and mobile devices 243 , 244 , and 245 are connected to base station 234 .
[0135] The device 211 includes a receiver 213, a transmitter 214, a processor 215, and a memory 217. The processor 215 is configured to: assemble the simulation models of the multiple components into different digital twin candidates of the telecommunications network; perform multiple simulations on each of the assembled different digital twin candidates to determine training samples; generate utility functions about one or more performance indicators for multiple digital twin candidates based on the training samples, the multiple digital twin candidates including the assembled different digital twin candidates; generate a cost function for the multiple digital twin candidates based on the training samples; and store information specifying the cost function and the utility function generated for the multiple digital twin candidates in the memory.
[0136] The device 201 includes a receiver 203, a transmitter 204, a processor 205, and a memory 207. The processor 205 is configured to: determine a minimum evaluation accuracy and / or determine a maximum evaluation duration with respect to one or more performance indicators; select a simulation model for each of a plurality of components of the digital twin based on at least one of the minimum evaluation accuracy and the maximum evaluation duration; assemble the selected simulation models into the digital twin; and perform simulation on the digital twin.
[0137] If the digital twin belongs to RAN 231, base stations 233 and 234 and other base stations can be represented as components of the digital twin. Antennas of base stations 233 and 234 and other base stations can also be represented as components of the digital twin. User mobility and propagation environment can also be represented as components of the digital twin. If the digital twin belongs to mobile communication network 200, core network components 223-224 can also be represented as components of the digital twin.
[0138] The processor 205 can be configured to select a simulation model for each of the multiple components of the digital twin based on at least one of a minimum evaluation accuracy and a maximum evaluation duration in the following manner: determine the number of the multiple components based on at least one of the minimum evaluation accuracy and the maximum evaluation duration, and select a simulation model for each of the multiple components.
[0139] Additionally or alternatively, if one or more component types are associated with multiple simulation models of different complexities, the processor 205 can be configured to select a simulation model for each of the multiple components of the digital twin based on at least one of a minimum evaluation accuracy and a maximum evaluation duration by (jointly) selecting a simulation model for each component of the one or more component types from multiple simulation models associated with the corresponding component type based on at least one of the minimum evaluation accuracy and the maximum evaluation duration.
[0140] exist Fig. 9In an embodiment, the processor 205 is configured to select a simulation model for each of the multiple components of the digital twin by applying the utility function generated by the device 211 to the minimum evaluation accuracy and / or by applying the cost function generated by the device 211 to the maximum evaluation duration. In an alternative embodiment, the processor 205 is configured to select a simulation model for each of the multiple components of the digital twin in a different manner.
[0141] exist Fig. 9 In an embodiment, device 201 and device 211 are separate devices. In an alternative embodiment, a single device performs the functions performed by device 201 and device 211. In another embodiment, the functions performed by device 201 are performed by multiple devices and / or the functions performed by device 211 are performed by multiple devices.
[0142] exist Fig. 9 In the illustrated embodiment, device 201 and device 211 include a processor 205 or 215. In alternative embodiments, device 201 and / or device 211 include multiple processors. For example, the processor can be a general-purpose processor (e.g., an Intel or AMD processor) or a special-purpose processor. For example, the processor can include multiple cores. For example, the processor can run a Unix-based operating system or a Windows operating system. For example, memory 205 and / or memory 215 can include solid-state memory (e.g., one or more solid-state disks (SSDs) made of flash memory) or one or more hard disks.
[0143] Receivers 203, 213 and transmitters 204, 214 may communicate with each other, or with other devices in, for example, a RAN or CN, using one or more communication technologies (wired or wireless). Receivers and transmitters may be combined into a transceiver. Devices 201 and 211 may include other components typical for a computer server or a network unit, such as a power supply.
[0144] In a typical example, the example optimization problem requires assembling a digital network twin that includes a mix of high-complexity and low-complexity components. The example necessitates periodic optimization of antenna downtilt angles in response to spatiotemporal variations in the offered traffic load.
[0145] In this typical example, the purpose of adjusting the antenna downtilt is to optimize the quality of service provided while ensuring a certain minimum degree of coverage (e.g. 99.5%) for a given load scenario. It is easy to understand that not only the absolute load level but also the spatial distribution of traffic will affect the optimal downtilt setting. For example, if all users are located close to the base station site, the optimal downtilt angle may be larger than if the users are evenly distributed in the area or even if the distribution is biased towards the cell edge. Since users tend to move over time (e.g. between working hours and non-working hours) and their communication needs also change over time (e.g. work-oriented emails during the day and HD video streaming at night), it makes sense that the optimal antenna tilt setting can also change over time.
[0146] The problem definition specifies a geographical area including, for example, one hundred base stations, specifies the antenna downtilt angle for each sector as a configuration parameter, and specifies a composite KPI as formulated above: optimize the quality of service (e.g., 10th percentile user throughput) at a coverage level exceeding, for example, 99.5%. In addition, the urgency level is marked as low and quantified as a maximum evaluation duration of up to several tens of minutes. This maximum evaluation duration can be included in the problem definition or specified separately. The latter is beneficial if the maximum evaluation duration can vary while the problem definition (of the rest) remains unchanged.
[0147] In this example, the digital twin required to properly model the network for this optimization problem includes simulation models for at least the deployed base station antennas, the propagation environment, and the spatial traffic distribution. For each base station antenna, a fairly accurate (high complexity) simulation model is required to realistically capture the antenna height, azimuth orientation, 3D radiation pattern, and transmit power, as all of these aspects significantly affect the coverage, the actually experienced SINR, and therefore the quality of service. It may be sufficient to include a large-scale representation of the propagation environment in the correspondingly selected (low complexity) simulation model, which covers in particular aspects of distance-based path loss, indoor penetration loss, and log-normal fading.
[0148] In this example, small-scale propagation aspects such as multipath fading can be excluded, since the associated small-scale variations are typically not addressed by adjusting antenna tilt settings, but rather by, for example, adaptive modulation and coding, transmit power control, and channel adaptive packet scheduling. Finally, an estimate of the spatial traffic distribution valid for the upcoming period is used. Since no user-specific level of detail is required (typically macroscopic accuracy up to 40×40m pixels is sufficient) and user mobility can be ignored, a simulation model of moderate complexity is considered appropriate for the spatial traffic distribution. It is worth noting that the digital twin assembled in this way deliberately omits aspects that may be relevant for other optimization studies, such as detailed modeling of various RRM mechanisms, such as packet scheduling, switching management, admission / congestion control.
[0149] The simulation involves considering multiple antenna tilt tuples. To find the optimal antenna downtilt, an initial tuple can be selected based on the current real-time configuration, and subsequent tuples can then be intelligently selected based on, for example, a gradient-based optimization strategy. All tuples are evaluated through the selected digital twin. Considering the composite KPI, the optimization strategy will track the best tilt tuple evaluated so far, and when the method reaches sufficient convergence, that tuple will be selected as the optimal solution and applied to the operational telecommunications network.
[0150] As mentioned above, the simulation model can be selected at the component type level or at the component level. For illustrative purposes, three simulation models with different levels of complexity can be distinguished for the component type propagation environment:
[0151] A low-complexity simulation model of the propagation environment can be as described above, including only large-scale aspects such as distance-based path loss, indoor penetration loss and log-normal fading. This model can be adapted for periodic optimization of antenna downtilt angles.
[0152] • A slightly more sophisticated simulation model (medium complexity) may also include small-scale multipath fading aspects, modeled as relatively simple time / frequency domain variations of the propagation gain around a large-scale mean. This model may be used to optimize the configuration of the packet scheduler.
[0153] Highly complex simulation models of the propagation environment involving all the above aspects and further including characterization of small-scale propagation aspects on a per transmit / receive antenna pair basis and even on a per-path basis, which may be necessary when optimizing physical layer algorithms, SU / MU-MIMO beamforming strategies, and corresponding multi-user grouping (co-)scheduling algorithms.
[0154] Fig.10 Describes the demonstration that can be performed as reference Figures 1 to 3 and Figures 5 and 6 A block diagram of an exemplary data processing system for describing the method.
[0155] like Fig.10 As shown, data processing system 300 may include at least one processor 302 coupled to memory element 304 via system bus 306. Therefore, the data processing system may store program code in memory element 304. Further, processor 302 may execute program code accessed from memory element 304 via system bus 306. In one aspect, the data processing system may be implemented as a computer suitable for storing and / or executing program code. However, it should be understood that data processing system 300 may be implemented in the form of any system including a processor and a memory capable of performing the functions described in this specification.
[0156] Memory element 304 may include one or more physical memory devices, such as local memory 308 and one or more mass storage devices 310. Local memory may refer to random access memory or other (multiple) non-persistent memory devices commonly used during the actual execution of program code. Mass storage devices may be implemented as hard drives or other persistent data storage devices. Processing system 300 may also include one or more cache memories (not shown) that provide temporary storage of at least some program code to reduce the number of times program code must be retrieved from mass storage device 310 during execution.
[0157] Input / output (I / O) devices, depicted as input device 312 and output device 314, may optionally be coupled to the data processing system. Examples of input devices may include, but are not limited to, a keyboard, a pointing device such as a mouse, etc. Examples of output devices may include, but are not limited to, a monitor or display, speakers, etc. Input devices and / or output devices may be coupled to the data processing system directly or through intervening I / O controllers.
[0158] In an embodiment, the input device and the output device may be implemented as a combined input / output device (in Fig.10 312 and output device 314). An example of such a combined device is a touch-sensitive display, sometimes referred to as a "touch screen display" or simply a "touch screen." In such an embodiment, input to the device can be provided by movement of a physical object (e.g., a stylus or a user's finger) on or near the touch screen display.
[0159] Network adapter 316 may also be coupled to data processing system 300 to enable it to couple to other systems, computer systems, remote network devices, and / or remote storage devices through intervening private or public networks. Network adapters may include data receivers for receiving data transmitted by the systems, devices, and / or networks to data processing system 300, and data transmitters for transmitting data from data processing system 300 to the systems, devices, and / or networks. Modems, cable modems, and Ethernet cards are examples of different types of network adapters that may be used with data processing system 300.
[0160] As in Fig.10 As depicted in FIG. 3 , memory element 304 may store application 318. In various embodiments, application 318 may be stored in local memory 308, one or more mass storage devices 310, or separate from local memory and mass storage devices. It should be understood that data processing system 300 may further execute an operating system (OS) that may facilitate the execution of application 318. Fig.10 318). Application 318 implemented in the form of executable program code may be executed by data processing system 300 (eg, by processor 302). In response to executing the application, data processing system 300 may be configured to perform one or more operations or method steps described herein.
[0161] Various embodiments of the present invention may be implemented as a program product for use with a computer system, wherein the (multiple) programs of the program product define the functions of the embodiments (including the methods described herein). In one embodiment, the (multiple) programs may be contained on various non-transient computer-readable storage media, wherein, as used herein, the expression "non-transient computer-readable storage medium" includes all computer-readable media, with the only exception of temporarily propagated signals. In another embodiment, the (multiple) programs may be contained on various transient computer-readable storage media. Illustrative computer-readable storage media include, but are not limited to: (i) non-writable storage media on which information is permanently stored (e.g., a read-only memory device within a computer, such as a CD-ROM disk, a ROM chip, or any type of solid-state non-volatile semiconductor memory readable by a CD-ROM drive); and (ii) writable storage media on which variable information is stored (e.g., a flash memory, a floppy disk within a floppy disk drive, or a hard disk drive, or any type of solid-state random access semiconductor memory). The computer program may run on the processor 302 described herein.
[0162] The terms used herein are for the purpose of describing specific embodiments only and are not intended to limit the present invention. As used herein, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that when the terms "comprises" and / or "comprising" are used in this specification, they specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.
[0163] The corresponding structures, materials, operations, and equivalents of all means or step-plus-function elements in the claims are intended to include structures, materials, or operations that perform functions in combination with other declared elements that are explicitly claimed. The description of the embodiments of the present invention has been presented for illustrative purposes, but the description is not intended to be exhaustive or to limit the implementation methods to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope of the present invention. The selection and description of the embodiments are intended to best explain the principles of the present invention and some practical applications, and when suitable for the specific purpose contemplated, enable other persons of ordinary skill in the art to understand the present invention for various embodiments with various modifications.
Claims
1. A system (201, 210) for performing simulation on a digital twin (1) of a telecommunications network (200), the system (201, 210) comprising at least one processor (205, 215), the at least one processor being configured to: - determining a minimum evaluation accuracy (66) with respect to one or more performance indicators and / or determining a maximum evaluation duration (67), - selecting a simulation model (13-15, 23-25, 33-35, 43-45) for each of a plurality of components (3-8) of the digital twin (1) based on at least one of the minimum evaluation accuracy and the maximum evaluation duration, - assembling the selected simulation model into the digital twin (1), and -Perform these simulations on the digital twin (1).
2. The system (201, 210) of claim 1, wherein: The at least one processor (205, 215) is configured to select the simulation model (13-15, 23-25, 33-35, 43-45) for each of the multiple components (3-8) of the digital twin (1) based on at least one of the minimum evaluation accuracy and the maximum evaluation duration in the following manner: determine the number of the multiple components based on at least one of the minimum evaluation accuracy and the maximum evaluation duration, and select the simulation model for each of the multiple components (3-8).
3. The system (201, 210) of claim 1 or 2, wherein: For at least one of the multiple components, the component type (11, 21, 31, 41) of the corresponding component is associated with multiple simulation models (13-15, 23-25, 33-35, 43-45) of different complexity, and wherein the at least one processor (205, 215) is configured to select the simulation model for each of the multiple components (3-8) of the digital twin (1) based on at least one of the minimum evaluation accuracy and the maximum evaluation duration in the following manner: based on at least one of the minimum evaluation accuracy and the maximum evaluation duration, at least one simulation model is selected from the multiple simulation models (13-15, 23-25, 33-35, 43-45) for the at least one component (3-8).
4. The system (201, 210) according to any one of the preceding claims, wherein: The at least one processor (205, 215) is configured to: - performing the simulations on the digital twin (1) to determine one or more network configuration settings of the telecommunications network (200), and - configuring the telecommunications network (200) using the one or more network configuration settings.
5. The system (201, 210) according to any one of the preceding claims, wherein: The at least one processor (205, 215) is configured to: - selecting a problem definition from a plurality of problem definitions, each of the plurality of problem definitions specifying one or more configuration parameters to be optimized and one or more corresponding performance indicators, - further selecting the simulation model for each of the plurality of components (3-8) of the digital twin (1) based on the selected problem definition, and -Select one or more configuration parameters for these simulations based on the selected problem definition.
6. The system of claim 5, wherein: The selected problem definition includes the minimum evaluation accuracy and / or the maximum evaluation duration.
7. The system (201, 210) according to any one of the preceding claims, wherein: The at least one processor (205, 215) is configured to: - Assemble simulation models into different digital twin candidates (70-73), - performing multiple simulations on each of the assembled different digital twin candidates (70-73) to determine training samples, - based on the training samples, generating a utility function (61) about the one or more performance indicators for a plurality of digital twin candidates (71-73), the plurality of digital twin candidates comprising assembled different digital twin candidates, - based on the training sample, generating a cost function (62) for the plurality of digital twin candidates (70-73), and - Selecting the simulation model for each of the plurality of components of the digital twin (1) by applying the utility function (61) to the minimum evaluation accuracy (66) and / or by applying the cost function (62) to the maximum evaluation duration (67).
8. The system (201, 210) of claim 7, wherein: The training samples include values of the one or more performance indicators, and the at least one processor (205, 215) is configured to generate the cost function (62) and / or the utility function (61) based on the values of the one or more performance indicators.
9. The system (201, 210) according to any one of the preceding claims, wherein: The at least one processor (205, 215) is configured to select the simulation model for each of the plurality of components (3-8) by selecting a digital twin candidate from a plurality of digital twin candidates (71-73) based on at least one of the minimum evaluation accuracy (66) and the maximum evaluation duration (67) and further based on expert knowledge.
10. The system (201, 210) according to any one of claims 1 to 9, wherein: The at least one processor (205, 215) is configured to select the simulation model for each of the multiple components (3-8) of the digital twin (1) based on the maximum evaluation duration (67), so that the simulations of the digital twin (1) have maximized evaluation accuracy without exceeding the maximum evaluation duration (67).
11. The system (201, 210) of any one of claims 1 to 9, wherein: The at least one processor (205, 215) is configured to select the simulation model for each of the multiple components of the digital twin (1) based on the minimum evaluation accuracy (66), so that the simulations of the digital twin (1) have a minimized evaluation duration while satisfying the minimum evaluation accuracy (66).
12. A system (210, 211) for generating a utility function and a cost function (61, 62) for a digital twin candidate of a telecommunications network, the system (210, 211) comprising at least one processor (205, 215) configured to: - assembling the simulation models (13-15, 23-25, 33-35, 43-45) into different digital twin candidates (70-73) of the telecommunication network (200), - performing multiple simulations on each of the assembled different digital twin candidates (70-73) to determine training samples, - based on these training samples, generating a utility function (61) about one or more performance indicators for a plurality of digital twin candidates (70-73), the plurality of digital twin candidates comprising assembled different digital twin candidates, - based on the training sample, generating a cost function (62) for the plurality of digital twin candidates (70-73), and -Storing in a memory (207, 217) information specifying the cost functions and utility functions (61, 62) generated for the plurality of digital twin candidates (70-73).
13. A computer-implemented method of performing simulation on a digital twin of a telecommunications network, the method comprising: - determining (101) a minimum evaluation accuracy with respect to one or more performance indicators and / or determining (101) a maximum evaluation duration; - selecting (103) a simulation model for each of a plurality of components of the digital twin based on at least one of the minimum evaluation accuracy and the maximum evaluation duration; - Assembling (105) the selected simulation model into the digital twin; as well as -Perform (107) these simulations on the digital twin.
14. A computer-implemented method for generating a utility function and a cost function for a digital twin candidate for a telecommunications network, the method comprising: - assembling (175) simulation models into different digital twin candidates of the telecommunication network; - performing (177) multiple simulations on each of the assembled different digital twin candidates to determine training samples, - based on the training samples, generating (179) a utility function on one or more performance indicators for a plurality of digital twin candidates, the plurality of digital twin candidates comprising assembled different digital twin candidates, - Based on the training samples, generating (181) a cost function for the plurality of digital twin candidates; as well as -Storing (183) in a memory information specifying the cost functions and utility functions generated for the plurality of digital twin candidates.
15. A computer program or computer program suite comprising at least one software code portion or a computer program product storing at least one software code portion, which software code portion is configured to perform the method according to claim 13 or 14 when run on a computer system.
