Assembling a digital twin of a telecommunications network

Abstract

A system for performing a simulation on a digital twin (1) of a telecommunications network is configured to determine a minimum assessment accuracy for one or more performance indicators and / or determine a maximum assessment period, select a simulation model (13-15, 23-25, 33-35, 43-45) for each of a plurality of elements (3-8) of the digital twin based on at least one of the minimum assessment accuracy and the maximum assessment period, assemble the selected simulation models (13, 14, 23, 33, 43) into a digital twin, and perform a simulation on the digital twin.

Description

[Technical field]

[0001] The present invention relates to a system for performing a simulation on a digital twin of a communication network.

[0002] The invention further relates to a method for performing a simulation on a digital twin of a telecommunications network.

[0003] The invention also relates to a computer program product enabling a computer system to carry out such a method. [Background technology]

[0004] The concepts of "cyber-physical fusion" and "digital twin" are being applied / developed in various domains including manufacturing, logistics and smart cities. The basic idea is to give a physical object a digital representation on which scenario-based experiments can be performed manually or automatically, for example to predict and analyze the object's behavior / reaction before selectively adapting the physical object itself.

[0005] There are various models / tools aimed at network planning, optimization and technology / feature development related to mobile networks. Widely used ones are e.g. wireless network planning tools, system and link level simulators, with different degrees of modeling accuracy / complexity and evaluation speed and therefore different targets of use. Link level simulators, on the one hand, model transmission, channel and reception behavior in great detail, and consequently treat only a single radio link in isolation. On the other hand, wireless network planning tools typically utilize a more or less simplified modeling of the network technology, allowing the assessment of relatively large parts of operational networks with many active users. Such tools are generally used by human experts to perform experiments in order to derive measures that are applied to the operational network at a later stage in a non-automated manner.

[0006] In the context of self-organizing networks (SONs), fully automated closed-loop control mechanisms have been developed and implemented, typically characterized by a continuous cycle of measurements, algorithmically processing the measurements, and applying the derived configuration changes to the operational network. The underlying algorithms can be if-then-else engineering rules or relatively simple artificial intelligence (AI) / machine learning (ML) based solutions, but are less evolved than solutions based on digital twins.

[0007] The paper “Digital twin for 5G and beyond”, HX Nguyen, R. Trestian, D. To and M. Tatipamula, IEEE Communications Magazine, vol. 59, no. 2, 2021, provides an overview of the idea of ​​digital twin in the context of mobile communication networks and how it can be used to optimize 5G networks. The authors present digital twin as including a virtual model that is self-learning in real time (using AI / ML techniques). The paper further describes digital twin as the integration of separate models that represent different parts of the network.

[0008] Ericsson's online blog "The future of digital twins: what will they mean for mobile networks?" (https: / / www.ericsson.com / en / blog / 2021 / 7 / future-digital-twins-in-mobile-networks) touches on the trade-off between the complexity, accuracy, and required computational resources of the simulations performed when applying the idea of ​​digital twins to mobile communication networks. The blog suggests mixing models of different complexity depending on the use case at hand, the network domain of interest, and the level of detail required. However, the blog does not disclose how this can be achieved.

[0009] Currently, there is no "one size fits all" solution for implementing digital twins in mobile networks or other telecommunications networks; the simulations performed by digital twins may be too slow in the first situation but not too slow in the second situation, or may be overly inaccurate in the first situation but not overly inaccurate in the second situation. Summary of the Invention [Problem to be solved by the invention]

[0010] A first object of the present invention is to provide a system capable of generating a digital twin that is suited to the situation at hand.

[0011] A second object of the present invention is to provide a method that can be used to generate a digital twin that is suited to the situation at hand. [Means for solving the problem]

[0012] In a first aspect of the invention, a system for performing a simulation on a digital twin of a telecommunications network includes at least one processor configured to determine a minimum assessment accuracy and / or determine a maximum assessment period for one or more performance metrics, select a simulation model for each of a plurality of elements of the digital twin based on at least one of the minimum assessment accuracy and the maximum assessment period, assemble the selected simulation models into a digital twin, and perform a simulation on the digital twin.

[0013] In the present system, in situations where assessment accuracy is important and / or assessment time period is not / less important, the digital twin may be assembled from complex elements and / or many elements, and in situations where assessment time period is important and / or assessment accuracy is not important, the digital twin may be assembled from less complex elements and / or fewer elements. In this way, the digital twin can be generated to suit the situation at hand. Examples of performance indicators are throughput, delay, coverage probability. Simulations are typically performed with different settings of configuration parameters. Simulation models are usually determined jointly, rather than for each element (separately). Each element has an element type. The telecommunications network may be, for example, a mobile communication network or a cable television network.

[0014] The at least one processor may be configured to select a simulation model for each of the plurality of elements of the digital twin based on at least one of a minimum assessment accuracy and a maximum assessment period by determining a quantity of the plurality of elements and selecting a simulation model for each of the plurality of elements based on at least one of a minimum assessment accuracy and a maximum assessment period. Thus, a digital twin may be assembled from more elements in situations where assessment accuracy is important and / or assessment period is not / less important. For example, a relatively larger number of elements may be selected for element types that are most relevant to the problem at hand.

[0015] Optionally, for at least one of the plurality of elements, the element type of each element may be associated with a plurality of simulation models having different degrees of complexity. The at least one processor may then be configured to select a simulation model for each of the plurality of elements of the digital twin based on at least one of the least assessment accuracy and the longest assessment period by (jointly) selecting at least one simulation model for the at least one element from the plurality of simulation models based on at least one of the least assessment accuracy and the longest assessment period. Thus, the digital twin may be assembled from more complex elements in situations where assessment accuracy is important and / or assessment period is not / less important.

[0016] The at least one processor may be configured to run simulations on the digital twin to determine one or more network configuration parameter settings for the telecommunications network and to configure the telecommunications network with the one or more network configuration parameter settings. This allows an operational telecommunications network to be configured with better network configuration parameter settings. These simulations may start from a current valid network configuration, or if this is not available, from e.g. default parameter settings, vendor recommended parameter settings, parameter settings within a range, or arbitrary parameter settings. The operational telecommunications network may be, for example, a trial network or a commercial network.

[0017] The one or more network configuration parameter settings may include, for example, antenna tilt settings and / or handover threshold settings and / or transmit power settings. The digital twin of the telecommunications network may additionally or alternatively be used to assess, for example, the addition of a new site at a time and location. In other words, a simulation may show that a new site would be beneficial.

[0018] The elements of the digital twin of a telecommunications network may, in the case of a mobile communications network, represent network elements of 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, equipment / 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 and one or more corresponding performance metrics to be optimized, select a simulation model for each of the plurality of elements of the digital twin further based on the selected problem definition, and select one or more configuration parameters for simulation from the selected problem definition. The problem definition may further specify a geographic scope and / or a technological scope. The selected problem definition may include a minimum assessment accuracy and / or a maximum assessment period.

[0020] In this way, an optimal number of elements of a digital twin and / or an optimal complexity of the simulation models of the elements may be selected for a particular problem, solving the problem of traditional "one size fits all" solutions where the digital twin is modeled in too much detail, resulting in unacceptably slow simulations for some problems, or insufficient detail to simulate with sufficient accuracy for other problems.

[0021] The defined problem may be an optimization problem or an evaluation problem. For example, the objective of an optimization problem may be the optimization of configuration parameters (e.g., tilt settings) in a mobile communication network, whereas the objective of an evaluation problem may be the assessment of the performance of a vendor-implemented feature with different settings of configuration parameters in a mobile communication network. The assessment of a feature essentially involves the optimization of the configuration parameters to quantify the benefit of the feature (i.e., the maximum performance gain achievable given the optimized configuration).

[0022] For example, in an optimization problem, the problem may be optimization of configuration parameters with respect to a single performance index or with respect to multiple performance indexes (e.g., a weighted average). The performance index may be a composite of multiple performance indexes. The optimization of the configuration parameters may be performed under certain constraints, e.g., the same constraints used to select the simulation models. For example, if a simulation model was selected for each of the multiple elements based on a minimum assessment accuracy, the optimization algorithm may use the same minimum assessment accuracy as a constraint.

[0023] The at least one processor may be configured to assemble the simulation models into different candidate digital twins, perform a plurality of simulations for each of the assembled different candidate digital twins to determine a training sample, generate a utility function for one or more performance metrics for the plurality of candidate digital twins including the assembled different candidate digital twins based on the training sample, generate a cost function for the plurality of candidate digital twins based on the training sample, and select a simulation model for each of the plurality of elements of the digital twin by applying the utility function to a lowest assessment accuracy and / or by applying the cost function to a longest assessment period. This is an advantageous method of selecting a simulation model for each of the plurality of elements of the digital twin based on at least one of a lowest assessment accuracy and a longest assessment period.

[0024] The utility function is also called the accuracy function. The cost function reflects the effort of the assessment, e.g., the number of evaluations required to determine the best / optimized settings of the configuration parameters multiplied by the computation time required to evaluate a single configuration of the digital twin. The computation time required to evaluate a single configuration is determined / learned for the system running the simulation. The utility and cost functions may be generated based on training samples, e.g., by using AI / ML techniques.

[0025] For example, in the case of an optimization problem, the optimization problem may be solved for each of multiple initial settings of configuration parameters for each candidate digital twin to determine the number of simulations / iterations that need to be performed for each initial configuration to arrive at an optimal / optimized configuration. These initial configuration parameter settings may be selected, for example, based on expert knowledge. By using AI / ML techniques, the utility and cost functions can be generated even without data on each possible (initial) setting of the configuration parameters.

[0026] It may be impossible to run multiple simulations for each candidate digital twin that can be assembled. The candidate digital twins from which the training samples are taken may thus be selected from the set of all possible digital twins that can be assembled based on expertise. This expertise may be related to the (domain of) the selected problem and / or the AI / ML techniques used.

[0027] The expert knowledge may be used at a later stage. At least one processor may be configured to select a simulation model for each of the multiple elements by selecting one digital twin candidate from the multiple digital twin candidates based on at least one of a (required) minimum assessment accuracy and / or a (acceptable) maximum assessment period, and further based on the expert knowledge. This may help in selecting the digital twin candidate that is best suited for the problem at hand in terms of the constraints.

[0028] In other words, a digital twin may be selected from a reduced set of all assemblable digital twins based on expertise. This expertise may be related to the (domain of) the selected problem and may be the same or (partially) different from the expertise used to select the digital twin candidates. The selection of a simulation model for each element of the digital twin may include, but is not required to include, selecting one of the digital twin candidates for which a training sample was obtained. For example, utility and cost function values ​​may be determined for unsimulated digital twin candidates, e.g. using AI / ML techniques.

[0029] The training samples may include values ​​of 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 values ​​of the one or more performance indicators. For example, the performance indicator values ​​obtained for a particular candidate digital twin may be compared to the performance indicator values ​​obtained for the most accurate candidate digital twin to generate a utility function.

[0030] The at least one processor may be configured to select a simulation model for each of the multiple elements of the digital twin based on a maximum (acceptable) assessment period such that a simulation for the digital twin has maximized assessment accuracy without exceeding the maximum (acceptable) assessment period. This is advantageous in cases where not exceeding the maximum (acceptable) assessment period is sufficient, but obtaining the shortest assessment period is not necessary, and it is more important to obtain the highest assessment accuracy subject to the assessment period constraint. This is a trade-off between assessment period and assessment accuracy.

[0031] Alternatively, the at least one processor may be configured to select a simulation model for each of the multiple elements of the digital twin based on a (required) minimum assessment accuracy such that a simulation for the digital twin has the shortest assessment time while meeting the (required) minimum assessment accuracy. This is advantageous if meeting the (required) minimum assessment accuracy is sufficient, but not necessary to obtain the highest assessment accuracy. This is another trade-off between assessment time and assessment accuracy.

[0032] In a second aspect of the present invention, a system for generating utility and cost functions for candidate digital twins of a telecommunications network includes at least one processor configured to assemble a simulation model into distinct candidate digital twins of the telecommunications network, perform a plurality of simulations for each of the distinct assembled candidate digital twins to determine a training sample, generate a utility function for one or more performance metrics for the plurality of candidate digital twins including the distinct assembled candidate digital twins based on the training sample, generate a cost function for the plurality of candidate digital twins based on the training sample, and store information specifying the generated cost and utility functions for the plurality of candidate digital twins in a memory.

[0033] In a third aspect of the invention, a computer-implemented method of performing a simulation on a digital twin of a telecommunications network includes determining a minimum assessment accuracy for one or more performance metrics and / or determining a maximum assessment duration, selecting a simulation model for each of a plurality of elements of the digital twin based on at least one of the minimum assessment accuracy and the maximum assessment duration, assembling the selected simulation models into the digital twin, and performing the simulation on the digital twin. The method may be performed by software running on a programmable device. The software may be provided as a computer program product.

[0034] In a fourth aspect of the invention, a computer-implemented method for generating utility and cost functions for candidate digital twins of a telecommunications network includes assembling a simulation model into distinct candidate digital twins of the telecommunications network, performing a plurality of simulations for each of the assembled distinct candidate digital twins to determine a training sample, generating utility functions for one or more performance indicators for the plurality of candidate digital twins including the assembled distinct candidate digital twins based on the training sample, generating cost functions for the plurality of candidate digital twins based on the training sample, and storing information specifying the generated cost and utility functions for the plurality of candidate digital twins in a memory. The method may be performed by software running on a programmable device. The software may be provided as a computer program product.

[0035] Further provided is a computer program for carrying out the methods described herein, and a non-transitory computer readable storage medium storing said computer program, which may for example be downloaded or uploaded to existing devices or may be stored at the time of manufacture of these systems.

[0036] The 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 to perform a simulation on a digital twin of a telecommunications network.

[0037] Executable actions include determining a minimum assessment accuracy and / or determining a maximum assessment period for one or more performance indicators, selecting a simulation model for each of a plurality of elements of the digital twin based on at least one of the minimum assessment accuracy and the maximum assessment period, assembling the selected simulation models into the digital twin, and running a simulation on the digital twin.

[0038] The 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 to generate utilities and cost functions for candidate digital twins of the telecommunications network.

[0039] The executable operations include assembling the simulation model into different candidate digital twins of the telecommunications network; running a plurality of simulations for each of the assembled different candidate digital twins to determine a training sample; generating a utility function for one or more performance metrics for the plurality of candidate digital twins including the assembled different candidate digital twins based on the training sample; generating a cost function for the plurality of candidate digital twins based on the training sample; and storing in memory information specifying the generated cost and utility functions for the plurality of candidate digital twins.

[0040] As will be appreciated by those skilled in the art, aspects of the present invention may be embodied as an apparatus, a method, or a computer program product. Accordingly, aspects of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining software and hardware aspects, all of which may be referred to generally herein as a "circuit," "module," or "system." Functions described in this disclosure may be implemented as an algorithm executed by a computer processor / microprocessor. Additionally, aspects of the present invention may take the form of a computer program product embodied in, e.g., stored on, one or more computer readable media having computer readable program code embodied therein.

[0041] Any combination of one or more computer readable media may be utilized. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. The computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination thereof. More specific examples of computer readable storage media include, but are not limited to, an electrical connection having one or more conductors, a portable computer diskette, a hard disk, a random access memory (RAM), a read only memory (ROM), an erasable programmable read only memory (EPROM or flash memory), an optical fiber, a portable compact disk read only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the context of the present invention, a computer readable storage medium may be any tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment.

[0042] A computer-readable signal medium may include a propagated data signal in which computer-readable program code is embodied, for example 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 is not a computer-readable storage medium but may be any computer-readable medium that can communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0043] The program code embodied in the computer readable medium may be transmitted using any suitable medium, including, but not limited to, wireless, wired, fiber optic, cable, RF, etc., or any suitable combination thereof. The computer program code for carrying out the operations of the aspects of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java™, Smalltalk, C++, and conventional procedural programming languages ​​such as the “C” programming language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server, as a stand-alone software package. In the latter scenario, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet Service Provider).

[0044] Aspects of the present invention are described below with reference to flow diagrams and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present invention. It will be understood that each block of the flow diagrams and / or block diagrams, and combinations of blocks in the flow diagrams and / or block diagrams, may be implemented by computer program instructions. These computer program instructions may be provided to a processor, in particular a microprocessor or central processing unit (CPU), of a general purpose computer, special purpose computer, or other programmable data processing apparatus to generate a machine such that the instructions, operating via the processor of the computer, other programmable data processing apparatus, or other device, generate means for performing the functions / operations specified in the block or blocks of the flow diagrams and / or block diagrams that are executed.

[0045] These computer program instructions, which can direct a computer, other programmable data processing device, or other device to function in a particular manner, can also be stored on a computer-readable medium, thereby constituting an article of manufacture including instructions that cause the instructions stored on the computer-readable medium to perform the functions / acts specified in the blocks or groups of blocks in the flow diagrams and / or block diagrams.

[0046] The computer program instructions may also be loaded into a computer, other programmable data processing apparatus, or other device to cause the computer, other programmable apparatus, or other device to perform a series of operational steps to generate a computer-implemented process such that the instructions operating on the computer or other programmable apparatus perform the functions / operations specified in the flow diagram and / or block diagram block or blocks.

[0047] The illustrated flow diagrams and block diagrams illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flow diagrams or block diagrams may represent a module, segment, or portion of code that comprises one or more executable instructions for implementing the specified logical function(s).

[0048] It should be noted that in some alternative embodiments, the functions described in the blocks may occur in a different order than that described in the figures. For example, two blocks shown in succession may in fact be executed substantially simultaneously, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. It should be noted that each block of the block diagrams and / or flow diagrams, and combinations of blocks in the block diagrams and / or flow diagrams, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or a combination of dedicated hardware and computer instructions.

[0049] These and other aspects of the invention will be apparent and elucidated with reference to the drawings, which are illustrative only and are not intended to be limiting. [Brief description of the drawings]

[0050] [Figure 1] FIG. 2 is a flow diagram of a first embodiment of a method for performing a simulation. [Diagram 2] FIG. 4 is a flow diagram of a second embodiment of a method for performing a simulation. [Diagram 3] FIG. 11 is a flow diagram of a third embodiment of a method for performing a simulation. [Figure 4] Figure 3 shows a digital twin being assembled in the manner described above. [Diagram 5] 1 shows different distinct phases of the method steps. [Figure 6] 13 shows a flow diagram of an embodiment of a method for generating cost and utility functions and a fourth embodiment of a method for performing a simulation. [Figure 7]Two examples of the use of exemplary cost and utility functions are given. [Figure 8] Two examples of the use of exemplary cost and utility functions are given. [Figure 9] FIG. 1 is a block diagram of an embodiment of a system. [Figure 10] FIG. 1 is a block diagram of an exemplary data processing system for implementing the methods of the present invention; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0051] The same elements in the drawings are designated with the same reference numbers.

[0052] A first embodiment of a computer-implemented method for performing a simulation on a digital twin of a telecommunications network is shown in Figure 1. Step 101 includes at least one of a) determining a (required) minimum assessment accuracy for one or more performance metrics and b) determining a (acceptable) maximum assessment period. Step 103 includes selecting a simulation model for each of a plurality of elements of the digital twin based on at least one of the minimum assessment accuracy and the maximum assessment period determined in step 101. Step 105 includes assembling the simulation models selected in step 103 into the digital twin.

[0053] Step 107 includes running a simulation on the digital twin assembled in step 105. The simulation may be performed, for example, to find one or more optimal configuration parameter settings. For example, a gradient-based optimization strategy may be used. The simulation model may be, for example, a black box in which specific parameter settings may be set / tuned.

[0054] Steps 101, 103, and 105 are part of the digital twin assembly phase 152, and step 107 is part of the digital twin utilization phase 153. According to one definition 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.

[0055] In the embodiment of FIG. 1, step 103 is performed by steps 121 and 123. Step 121 includes determining a quantity of a plurality of elements based on at least one of a minimum assessment accuracy and a maximum assessment period. In the embodiment of FIG. 1, the quantity of a plurality of elements is determined in step 121 by determining a quantity of elements for each element type. Optionally, one or more (but not all in this embodiment) element types have a fixed quantity of elements. The quantity of elements for each other element type is (jointly) determined based on at least one of a minimum assessment accuracy and a maximum assessment period. Step 121 may include, but is not required to include, adding up the quantities of elements for each element type.

[0056] Step 123 includes selecting a simulation model for each of the elements. In the embodiment of FIG. 1, one simulation model is associated with each element type. As a result, the same simulation model is selected for each element of the same element type. As an example, a digital twin of a radio access network may use element types A) base station, B) antenna, C) propagation environment, and D) user mobility. If two elements of type A, four elements of type B, one element of type C, and one element of type D are determined in step 121, eight simulation models are selected in step 123, of which two simulation models of type A are identical and four simulation models of type B are identical.

[0057] A second embodiment of a computer-implemented method for performing a simulation on a digital twin of a telecommunications network is shown in Figure 2. The method of Figure 2 also includes steps 101, 105, and 107, similar to the method of Figure 1. However, in the embodiment of Figure 2, step 103 is performed by steps 131 and 133, and for at least one of the plurality of elements, a plurality of simulation models of different complexities are associated with the element type of each element.

[0058] Step 131 includes determining a quantity of a number of elements. In the embodiment of Figure 2, a fixed quantity of elements is determined for all element types, i.e., the quantity of elements does not depend on the minimum assessment accuracy or the maximum assessment period. The quantity of elements may still depend on the selected problem definition, as will be described with respect to Figure 6. In a variation of the embodiment of Figure 2, step 131 is omitted.

[0059] Step 133 includes (jointly) selecting a simulation model for each of the elements having an element type associated with multiple simulation models based on the lowest assessment accuracy and / or the longest assessment period. Step 133 further includes selecting a simulation model for each of the elements having an element type if there is an element type associated with only one simulation model. If the same simulation model of the same element type is selected for each element, step 131 may alternatively be performed after step 133 or in parallel with step 133.

[0060] A third embodiment of a computer-implemented method for performing a simulation on a digital twin of a telecommunications network is shown in Figure 3. The method of Figure 3 also includes steps 101, 105, and 107, similar to the methods of Figures 1 and 2. However, in the embodiment of Figure 3, step 103 is performed by step 121 of Figure 1 and step 133 of Figure 2.

[0061] FIG. 4 shows a digital twin assembled in the manner of FIG. 3. FIG. 4 shows four element types 11, 21, 31, and 41. In the example of FIG. 4, each element type is associated with three different simulation models with different complexity levels (low, medium, high). These simulation models are stored in a repository with well-defined inter-element interfaces that support adapted assembly of simulation models of digital twin element types. The digital twin of FIG. 4 is only an example, and may have more or less than four element types and / or more or less than three complexity levels, for example.

[0062] The element type 11 is associated with a low complexity simulation model 13, a medium complexity simulation model 14, and a high complexity simulation model 15. The element type 21 is associated with a low complexity simulation model 23, a medium complexity simulation model 24, and a high complexity simulation model 25. The element type 31 is associated with a low complexity simulation model 33, a medium complexity simulation model 34, and a high complexity simulation model 35. The element type 41 is associated with a low complexity simulation model 43, a medium complexity simulation model 44, and a high complexity simulation model 45.

[0063] 4, in step 121 of FIG 3, two elements 3 and 4 are determined for element type 11, two elements 5 and 6 are determined for element type 21, one element 7 is determined for element type 31, and one element 8 is determined for element type 41. These six elements 3 to 8 are shown as elements of the digital twin 1 in FIG 4.

[0064] In the example of Fig. 4, in step 133 of Fig. 3, simulation models with different degrees of complexity are selected for two elements 3 and 4 of element type 11, i.e., low complexity simulation model 13 and medium complexity simulation model 14, and the same simulation model is selected for two elements 5 and 6 of element type 21, i.e., low complexity simulation model 23. Furthermore, a low complexity simulation model 33 is selected for a single element 7 of element type 31, and a low complexity simulation model 43 is selected for a single element 8 of element type 41. These simulation models are shown assembled into the digital twin 1 in Fig. 4.

[0065] In the embodiment of Figures 1-3, the simulation model is selected based only on the lowest assessment accuracy and / or the longest assessment period. In an alternative embodiment, as described with respect to Figure 6, the simulation model may be selected based on expert knowledge of the problem given. For example, the expert knowledge may be used to determine, for each element type, a suitable minimum and / or maximum number of elements and / or a suitable minimum and / or maximum complexity of the simulation model. For example, the expert knowledge may specify that for a given problem, a high complexity user mobility model is required and / or at least 57 base stations are required.

[0066] Illustrative examples of digital twin elements with different levels of complexity include: · Base Station: On the one hand a fully functional replica of a base station (possibly provided by the equipment vendor) may be used, whereas on the other hand the base station may be reduced to only a few key radio resource management mechanisms in combination. Antennas: If the problem of interest concerns the optimization of beamforming parameters, a detailed model of the deployed antenna array may be required, including, among other things, the number of antenna elements, the subarray structure and element-level antenna diagrams, whereas for other problems it is sufficient to reduce the antenna array to just a single effective transmit / receive antenna with a suitably configured antenna diagram. Propagation environment: For problems involving beamforming or channel adaptive packet scheduling, it is important to properly incorporate multipath fading in the propagation model, whereas for other problems it is sufficient to leave out multipath fading entirely and model only the path loss and possibly shadow fading. User mobility: When optimizing handover or dynamic beam steering parameters, it may be important to consider detailed user-level mobility models. For other problems, a macroscopic model of e.g. traffic flows between industrial and residential areas may be sufficient, for other problems user mobility can be ignored completely.

[0067] Other examples include user equipment, time-space traffic characteristics, and various core network functions and interfaces.

[0068] As mentioned with respect to Fig. 1, Fig. 1 shows a digital twin assembly phase 152, and a digital twin utilization phase 153. Fig. 5 further shows a preparation phase 151 prior to the assembly phase 152. The preparation phase 151 is typically an offline phase, whereas the assembly phase 152 and utilization phase 153 are online phases. The preparation phase 151 is directed to the development of simulation models of element types of the digital twin of the telecommunication network of interest, possibly including measurements made on the telecommunication network and AI / ML methodologies of the modeling itself. In one embodiment, the preparation phase 151 is also responsible for the generation of various utility or cost functions (e.g., for assessment accuracy and assessment duration), see e.g. Fig. 6.

[0069] Depending on the context, the telecommunications network may keep changing, so the preparation phase 151 may be a continuous process of developing / adjusting simulation models of element types. The assembly phase 152 deals with the adapted assembly of the digital twins by selecting the relevant models built in the preparation phase 151. In the exploitation phase 153, the assembled digital twins are exploited in simulations, for example to assess the impact of candidate configurations / actions and to apply the selected configurations / actions to the twinned telecommunications network.

[0070] In the embodiment of Figures 1-3, the utilization phase 153 includes step 107. Step 107 includes running a simulation on the digital twin assembled in step 105. In a variation of the embodiment of Figures 1-3, step 107 is performed by step 141, and the utilization phase 153 further includes step 143. This is shown in Figure 5.

[0071] Step 141 includes running a simulation on the digital twin assembled in step 105 to determine one or more network configuration parameter settings of the communication 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 twinned telecommunication network with the one or more network configuration settings determined in step 141.

[0072] One embodiment of a computer-implemented method for generating cost and utility functions and a fourth embodiment of a computer-implemented method for running a simulation on a digital twin of a telecommunications network are shown in Figure 6. The method for generating cost and utility functions is performed in a preparation phase 151. The method for running a simulation on a digital twin, including steps 101-107, is performed in an assembly phase 152 and a utilization phase 153.

[0073] As explained above, the preparation phase 151 covers the development of simulation models of the element types of the digital twin of the telecommunications network of interest, possibly including AI / ML methodologies of measurements made in the telecommunications network and the modeling itself. This may be performed with conventional techniques. It is assumed that a repository of simulation models of the element types of the digital twin with well-specified inter-element interfaces supporting assembly is available.

[0074] The preparation phase 151 includes steps 171, 173, 175, 177, 178, 179, 181, 183, and 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 corresponding one or more performance indicators. One or more performance indicators that need to reach minimum and / or maximum values ​​may optionally be included in the problem definition (as constraints). Examples of performance indicators are throughput and delay.

[0075] Examples of configuration parameters have been described with respect to Figure 5. The problem definition may further specify a geographical scope and / or a technological scope. The problem definition may further specify a minimum assessment accuracy (required) and / or a maximum assessment period (allowed), although if not, this may be entered separately in the assembly phase 152.

[0076] The problem definition may, for example, specify an optimization or evaluation problem. In either case, simulations are performed to find one or more optimal / optimized configuration parameter settings with respect to a performance index (typically subject to certain constraints). The objective of solving an optimization problem is to find one or more optimal / optimized configuration parameter settings. The objective of solving an evaluation problem is to find values ​​of the performance index that correspond to these one or more optimized / optimized configuration parameter settings.

[0077] In the assembly phase 152, a problem may be triggered periodically as part of an automated network optimization loop, or due to a base station failure, or due to an accidental event such as a network engineer triggering an evaluation of a vendor-implemented feature in the mobile operator's network. If a problem occurs in the assembly phase 152, a corresponding problem definition is selected from a plurality of problem definitions, e.g., from a list of problem definitions.

[0078] In terms of the target performance metric, a well-specified optimization / evaluation problem should preferably include a clear definition of what needs to be optimized / evaluated, e.g., simply a single KPI (e.g., coverage probability), multiple KPIs (e.g., K X , K Y ), i.e., K=α X K X +β Y K Y ,(e.g. 0.8 x drop probability + 0.2 x handover ping-pong ratio), or a combination of optimization KPIs and conditional KPIs, i.e.

number

number

[0079] In a first iteration of step 173, step 173 includes selecting a simulation model for each of a plurality of elements of the first candidate digital twin. Step 173 may be similar to step 103 of FIGS. 1-3. Thus, a desired quantity of elements may be determined in step 173 (e.g., per element type) and / or a simulation model having a desired degree of complexity may be determined in step 173 (e.g., per element or element type). The desired quantity of elements may depend on the degree of complexity of the selected simulation model. In a first iteration of step 175, step 175 includes assembling the simulation models of the plurality of elements into a first candidate digital twin. Step 175 may be similar to step 105 of FIGS. 1-3.

[0080] Each (candidate) digital twin is a vector n It can be expressed as the vector n indicates, for each twin element type and associated complexity level available in the repository, how many elements from that repository are included in the digital twin (e.g. n =#high complexity level BS, #medium complexity level BS, #low complexity level BS, #high complexity level antenna, #medium complexity level antenna, #low complexity level antenna, etc.). Some of these may be set to 0.

[0081] Step 177 includes performing multiple simulations on the candidate digital twins assembled in step 175 to determine a training sample. n In, iterative simulations are run at various settings of one or more configuration parameters (e.g., a range of gradient settings). The training samples typically include values ​​of one or more performance indices. n For this, step 177 may include the following two substeps. Determine one or more configuration parameter settings (e.g. a range of gradient settings to evaluate a particular problem). Expert knowledge can be used to select one or more configuration parameter settings, i.e. to select candidate configurations. This substep may include the selection of an initial setup. Given n and perform actual simulations for the candidate configurations. n Different trials are performed for x, y ... n If multiple trials are performed starting from multiple different initial configurations, the average total time per trial is n The individual outputs of this sub-stage may be stored / output in association with the training samples (e.g., [problem, n , KPI name]->[KPI value, total time]). If the optimized configuration is at the edge of the range of candidate configurations, the range can be adjusted to cover candidate configurations beyond that edge. In this case, the first and second sub-phases can be iterated and another candidate / initial configuration selected in the first sub-phase. An example of an optimization algorithm that can be used is gradient-based optimization in a multi-dimensional space. The individual simulation models for each element type are usually black boxes.

[0082] Step 178 includes determining whether simulations need to be run for additional digital twin candidates, i.e., whether simulations have been run for a sufficient number of digital twin candidates. Expert knowledge is preferably used to constrain the set of digital twin candidates from which digital twin candidates are selected in step 173. In this case, step 178 includes determining whether a sufficient number of digital twin candidates have been selected from the set to generate suitable cost and utility functions with reasonable effort.

[0083] If simulations need to be performed for additional candidate digital twins, step 173 is repeated and the method proceeds as shown in Figure 6. In the next iteration of step 173, step 173 includes selecting a simulation model for each of the plurality of elements of the next candidate digital twin. In the next iteration of step 175, step 175 includes assembling the simulation models of the plurality of elements into the next candidate digital twin.

[0084] Steps 179 and 181 are performed after step 178. Step 179 includes generating a utility function for one or more performance indicators based on the training samples determined in step 177 for a plurality of digital twin candidates. The plurality of digital twin candidates includes each digital twin candidate assembled in step 175. The utility function reflects an estimated accuracy for one or more performance indicators. Step 181 includes generating a cost function for the plurality of digital twin candidates based on the training samples determined in step 177. The cost function reflects an experienced and optionally estimated assessment effort. Steps 179 and 181 may include, for example, using machine learning techniques to estimate assessment duration and accuracy for twin candidates for which no training samples have been obtained. In alternative embodiments, only utility functions are generated or only cost functions are generated.

[0085] Typically, the utility function is determined based on the values ​​of one or more performance indicators contained in the training samples. For example, the utility function is generated by first determining the candidate digital twins using the candidate digital twin that yields the most accurate performance results, i.e., the full swing digital twin. The full swing digital twin considers the most complex simulation model available for each element type, and considers the entire network or a subset thereof that yields the same assessment accuracy. The utility function may be determined by normalization such that the function result is 1 for the candidate digital twin and 0-1 for other candidate digital twins. The resulting value of the utility function for a particular candidate digital twin reflects the utility of that candidate digital twin relative to the candidate digital twin with the highest utility (i.e., a result value of 1).

[0086] The cost function can be generated, for example, by determining, for each candidate digital twin, the average or maximum total time it takes to determine an optimal / optimized configuration (e.g., the average or maximum number of iterations of the optimization algorithm using different initial configurations), or by determining, for each candidate digital twin, the average or maximum number of simulations / iterations it takes to determine an optimized configuration and multiplying that average or maximum number of simulations / iterations by the computational time (e.g., in seconds) it takes to run one simulation / iteration.

[0087] For each configuration considered, sufficient replications with different random seeds are preferably performed to obtain sufficient statistical confidence in the performance results obtained. Thus, each iteration may include one or more replications. Examples of utility and cost functions are shown in Figures 7 and 8. Step 183 includes storing in memory information specifying the cost and utility functions generated for the problem definition selected in step 171.

[0088] Step 187 includes checking whether cost and utility functions should be generated for further problem definitions. If steps 173-183 have been performed for all of the plurality of problem definitions, then preparation phase 151 is complete (but may not be repeated later). If steps 173-183 have not been performed for all of the plurality of problem definitions, then step 171 is repeated and the method proceeds as shown in Figure 6. In the next iteration of step 171, step 171 includes selecting a next problem definition from the plurality of problem definitions.

[0089] Machine learning can be used to determine the resulting cost and utility function values ​​associated with candidate digital twins for which no simulation has been performed. It is also possible to use training samples from one problem to learn the cost and utility functions for another problem. In that case, steps 179, 181, and 183 may be performed after step 187 rather than before. In either case, cost and utility functions are generated for each problem from multiple, e.g., a list of potential problems.

[0090] As explained above, the assembly phase 152 handles the adapted assembly of the digital twins by selecting the relevant models built in the preparation phase 151. In the embodiment of Figure 6, relevant utility and cost functions are selected in the assembly phase 152 to determine the digital twins when a real problem occurs.

[0091] The assembly phase 152 includes steps 191, 193, 195, 101, 103 and 105. Step 191 includes a step of selecting a problem definition from a plurality of problem definitions. In the assembly phase 152, the selected problem definition should correspond to a problem occurring in an operational telecommunications network. Therefore, if the problem definition is not selected completely automatically by the system but based on user input, the user should not select a problem definition that does not correspond to the occurring problem. The operational telecommunications network may be, for example, a test network or a production network. Step 193 includes a step of reading from memory the cost and utility functions generated in relation to the problem definition selected in step 191.

[0092] Step 101 includes determining a (required) minimum assessment accuracy and / or a (acceptable) maximum assessment duration for one or more performance indicators specified in the selected problem definition. In the embodiment of Figure 6, the minimum assessment accuracy and / or the maximum assessment duration are determined separately in step 101 and are not specified in the problem definition. In an alternative embodiment, the problem definition includes a (required) minimum assessment accuracy and / or a (acceptable) maximum assessment duration.

[0093] Step 103 is executed after steps 193 and 101 are executed. Step 103 is executed by step 195. Step 195 includes selecting a simulation model for each of a plurality of elements of the digital twin by performing at least one of the steps of: a) applying a utility function corresponding to the problem definition retrieved in step 193 to the lowest assessment accuracy determined in step 101, and b) applying a cost function corresponding to the problem definition retrieved in step 193 to the longest assessment period determined in step 101. Step 195 can be considered as selecting a digital twin from a plurality of digital twin candidates based on at least one of the lowest assessment accuracy and the longest assessment period and further based on the selected problem definition.

[0094] The digital twin candidates may differ from each other by having different quantities of elements and / or by having simulation models of different complexity, as described with respect to figures 1-3. Step 195 may comprise a first sub-step of selecting a subset of the plurality of digital twin candidates and a second sub-step of selecting the digital twin from the subset. The subset may for example be selected based on expert knowledge. For example, for a gradient optimization problem, only BSs with low complexity need to be considered. Information facilitating the selection of the subset (and information representative of the expert knowledge) may for example be included in the problem definition.

[0095] The plurality of candidate digital twins may include all possible candidate digital twins, or may even be infinite, in which case the plurality of candidate digital twins may include candidates digital twins for which no training samples were obtained in step 177, but for which utility and cost values ​​can be derived / learned from utility and cost values ​​determined for other candidate digital twins. Alternatively, the plurality of candidate digital twins may only include candidates digital twins for which training samples were obtained in step 177.

[0096] As mentioned above, selecting an optimal level of complexity and / or an optimal number of elements per element type is important for adapting the digital twin, and this selection can be made based on a minimum assessment accuracy and / or a maximum assessment period.

[0097] The actual assessment period usually depends on the available computational resources. Since higher complexity requires more computational resources, e.g., processing power or memory resources, the availability of such resources may also affect the feasibility of solving a given optimization problem within a given deadline and for a given complexity of the twin. As a result, computational resource constraints may constrain the allowed complexity. It may be assumed that the exploitation phase 153 is performed with the same computational resources as the preparation phase 151, and thus the cost function can be used without adjustment or how to convert time to actual computational resources is known.

[0098] The (acceptable) maximum assessment period usually 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, a sub-optimal solution may be found as long as a quick (fast but crude) optimal solution can be found and thus the complexity of the twin is reduced. Alternatively, a process that periodically updates antenna tilt and power settings in response to spatio-temporal changes in the offered traffic load does not require an urgent solution and therefore a (near-optimal) solution can be achieved using a more complex twin.

[0099] A trade-off may be required between the quantity of elements selected per element type and the complexity of the elements selected per element type. This trade-off usually depends on the optimization problem target and may be made with the aid of expert knowledge. Some optimization problems are inherently geographically local in nature, e.g. cell-specific configuration of packet scheduling parameters. As a result, twin-based experiments to derive optimal parameter settings can be performed for single-cell scenarios, allowing the use of detailed twins in a reasonable time. Alternatively, the antenna tilt optimization problem described above inherently requires a larger (regional) scope, since tilt settings affect inter-cell interference and jointly determine coverage. As a result, the complexity of the twin should be lower than the previous (local) optimization example, so that per-antenna tilt settings can be extensively evaluated in a reasonable time.

[0100] Besides being limited in geographic scope, the problem may also be characterized as being limited in technical / network scope, for example to include only the RAN or more sophisticated RAN / CN scope, or to include only a limited set of protocol layers. Similar to limited geographic scope, here too the scope may impact the modeling scope and therefore the degree of twin detail required.

[0101] Step 105 includes assembling the simulation model selected in step 103 (e.g., with respect to the quantity of elements determined in step 103 and / or the complexity determined in step 103), i.e., the simulation model associated with the digital twin selected from the plurality of candidate digital twins, to obtain a digital twin. In the embodiment of FIG. 6, the candidate digital twin assembled in step 175 is not stored in memory. In an alternative embodiment, if the digital twin selected from the plurality of candidate digital twins has already been assembled and stored in memory, e.g., in step 175, the selected digital twin may be retrieved from memory.

[0102] The adapted digital twin assembled in the assembly phase 152 is then applied in an exploitation phase 153 in a collection of problem-specific simulations to derive one or more optimal / optimized configuration parameter settings (e.g., tilt settings) that may be sequentially applied to an operational telecommunications network, e.g., a test telecommunications network or a commercial telecommunications network.

[0103] The exploitation phase 153 includes steps 197 and 107. Step 197 includes selecting one or more configuration parameters specified in the problem definition selected in step 191. Step 107 runs simulations with different settings of the one or more configuration parameters selected in step 197 on the digital twin assembled in step 105, e.g. starting from a current valid network configuration, or if this is not available, e.g. starting from default parameter settings, vendor recommended parameter settings, parameter settings within a range, or arbitrary parameter settings.

[0104] The simulation performed in step 107 is similar to the simulation performed in step 177, but for one problem and one assembled digital twin, i.e., the selected candidate digital twin. If the (acceptable) maximum duration allows, the simulation can be performed with multiple initializations. For example, new initializations can be suggested based on knowledge gained from a given set of simulations with one or more initializations. The optimization algorithm and stopping criteria used in step 177 may also be used in step 107. As mentioned above, one example of an optimization algorithm that can be used is gradient-based optimization in a multidimensional space. Optionally, step 107 is performed by step 141, as described with respect to FIG. 5, and step 107 is followed by step 143.

[0105] Two use cases of exemplary cost and utility functions are shown in Figures 7 and 8. Although there may be situations where only a cost function or only a utility function is required, typically the selection of a digital twin from multiple candidate digital twins involves a trade-off between the accuracy of the results and the assessment effort (i.e., the time required to determine the best / optimized configuration). As discussed with respect to Figure 6, based on (e.g., periodic) offline experiments with different candidate digital twins, relevant utility or cost functions (e.g., related to estimated assessment accuracy and estimated assessment effort / time) may be generated, e.g., by machine learning.

[0106] Figures 7 and 8 show the utility function U P ( n )61 and cost function C P ( n )62. The utility function U P ( n ) 61, also called the accuracy function, represents the relative closeness of the performance achieved for each twin option for a particular problem P to the full-swing (most realistic, closest to the real-world object, and assumed to be the ground truth) model of the twin. This curve 61 thus converges to "1" as the twin options considered become as complex as the full-swing twin, and in fact may reach the "1" level for twin candidates that are less complex than the full-swing model, depending on the problem being addressed. For each element type, the full-swing digital twin 73 considers the most complex element model available with a sufficiently large number of elements. Ideally, the full-swing digital twin 73 should consider the entire network. However, in practice, a subset of the entire network may suffice.

[0107] Cost function C P ( n )62 is also called the assessment effort function and is a digital twin n Denotes the estimated effort (e.g., in seconds) required to determine the optimal / optimized configuration parameter settings for the cost function C P ( n )62 is E P ( n ) and T P ( n ) may be multiplied by E P ( n ) is the expected number of configurations that need to be simulated to determine the optimal configuration parameter settings (i.e., the number of gradient vectors, i.e., iterations, that need to be simulated as part of the optimization algorithm to determine the optimal / optimized configuration parameter settings, which typically increases with the number of BSs included in the twin). T P ( n )teeth, nis the computation time required to accurately simulate a single configuration of the digital twin given by

[0108] As mentioned above, the function U P ( n ), E P ( n ), T P ( n ) depends on the problem (P). P ( n Varying U depending on the problem can lead to a good estimation of the expected number of simulations required, since it is not averaged over all potential problems (e.g. the number of possible gradient settings differs from the number of possible scheduling parameter settings, and therefore the optimization will require a different number of simulations / iterations). For a particular problem, U P ( n ) can be obtained by observing or estimating (e.g., by using interpolation or AI / ML models) one or more realized (potentially composite) performance indicators, and T P ( n ) can be obtained by calculating the required simulation time for a target (potentially composite) KPI.

[0109] If a periodic or event-driven problem triggers the use of a digital twin, the optimal twin option may be selected based on the minimum assessment accuracy (required) and the maximum assessment period (acceptable). To find the optimal digital twin, the accuracy of the results can be traded off against the assessment effort in different ways. For example, the following options can be considered: I C P ( n )≦C MAX According to U P ( n ), or II.U P ( n )≧U MIN According to C P ( n ) is minimized.

[0110] As described with respect to step 195 of FIG. 6, first a subset of multiple digital twin candidates may be selected based on expertise, and then a digital twin may be selected from the subset. This subset may be selected, for example, based on problem expertise. For example, for a gradient optimization problem, only BSs with low complexity need to be considered. This expertise may be the same as the expertise used in the preparation phase 151 or may be (partially) different. For example, the expertise used in the preparation phase 151 may include machine learning knowledge in addition to problem knowledge.

[0111] Option II is shown in FIG. 7. FIG. 7 shows the minimum assessment accuracy U MIN FIG. 1 illustrates a method for selecting a simulation model for each of a plurality of elements of a digital twin based on a minimum required assessment accuracy such that a simulation for the digital twin is performed in a shortest assessment period while satisfying the above. As explained above, the quantity of the plurality of elements may be selected as part of the selection of the simulation model for each of the plurality of elements and may not be given a priori.

[0112] The digital twin candidate with the shortest assessment period under the above conditions is digital twin candidate 71, and is therefore selected. The highest assessment accuracy (i.e., 1) is represented by line 65. Full-swing digital twin candidate 73 has the highest assessment accuracy. Curve 61 increases relatively quickly, and relatively many digital twin candidates have assessment accuracies that are 1 or close to 1. The assessment accuracy of digital twin candidate 72 is approximately the same as full-swing digital twin candidate 73. Full-swing digital twin candidate 73 and digital twin candidate 72 have assessment accuracies that are at least as high as the minimum assessment accuracy U MIN 70 Digital Twin candidates have U P ( n ) is U MIN Since it is not more than that, the condition is not met.

[0113] Option I is shown in Figure 8. Figure 8 shows that the simulation for the digital twin is performed over the maximum assessment period C MAX 10 illustrates that a simulation model is selected for each of the plurality of elements of the digital twin based on a longest allowable assessment period to have a maximized assessment accuracy without exceeding 100%. As explained above, the quantity of the plurality of elements may be selected as part of the selection of the simulation model for each of the plurality of elements and may not be given a priori.

[0114] Under the above conditions, the digital twin candidate with the highest assessment accuracy is digital twin candidate 72, and therefore has been selected. P ( n ) exceeds the longest evaluation period for full-swing digital twin candidate 73, so full-swing digital twin candidate 73 does not meet the condition. Digital twin candidates 70 and 71 meet the condition, i.e., C P ( n ) is C MAX , but does not have maximized assessment accuracy.

[0115] In the example of Figures 7 and 8, only four digital twin candidates 70-73 are considered, but the curves 61 and 62 represent more than these four digital twin candidates. The fact that the digital twin candidates 70-73 are considered does not necessarily mean that training samples of these (all) digital twin candidates were obtained in the preparation phase. Any option other than options I and II is possible. Minimizing the assessment period or maximizing the assessment accuracy is not required.

[0116] 9 is a block diagram of an embodiment of a system for performing simulations on a digital twin of a communication network and a system for generating utility and cost functions for a candidate digital twin of a communication network. The system for generating utility and cost functions includes at least an apparatus 211 and may further include, for example, an apparatus 201. The system for performing simulations on a digital twin of a telecommunication network includes at least an apparatus 201 and may further include, for example, an apparatus 211. The system 210 includes both the apparatus 201 and the apparatus 211. In the example of FIG. 9, the telecommunication network is a mobile communication network 200. The mobile communication network 200 includes, among other things, a radio access network (RAN) 231 and a core network (CN) 221.

[0117] The RAN 231 includes base stations 233 and 234. The mobile communication network 200 may be a 5G network, the RAN 231 may be a 5G new wireless RAN, and the base stations 233 and 234 may be, for example, 5GgNodeB base stations. Each of the base stations 233 and 234 may include multiple distributed units that share a common centralized unit in a centralized RAN (C-RAN) architecture. The CN 221 includes core network elements 223, 224, and 225.

[0118] Apparatus 201 is used online to manage RAN 231 and CN 221 and may be part of a management network of mobile communication network 200. Apparatus 211 is used offline to generate utility and cost functions. In the example of Figure 9, mobile devices 241, 242 are connected to base station 233, and mobile devices 243, 244 and 245 are connected to base station 234.

[0119] Apparatus 211 includes receiver 213, transmitter 214, processor 215, and memory 217. Processor 215 is configured to assemble simulation models of the multiple elements into distinct candidate digital twins of the telecommunications network, perform multiple simulations for each of the assembled distinct candidate digital twins to determine a training sample, generate a utility function for one or more performance metrics for the multiple candidate digital twins including the assembled distinct candidate digital twins based on the training sample, generate a cost function for the multiple candidate digital twins based on the training sample, and store information specifying the generated cost and utility functions for the multiple candidate digital twins in memory.

[0120] Apparatus 201 includes a receiver 203, a transmitter 204, a processor 205, and a memory 207. Processor 205 is configured to determine a minimum assessment accuracy for one or more performance metrics and / or determine a maximum assessment duration, select a simulation model for each of a plurality of elements of the digital twin based on at least one of the minimum assessment accuracy and the maximum assessment duration, assemble the selected simulation models into a digital twin, and perform a simulation on the digital twin.

[0121] If the digital twin is of the RAN 231, the base stations 233, 234 and other base stations can be represented as elements of the digital twin. The antennas of the base stations 233, 234 and other base stations can also be represented as elements of the digital twin. The user mobility and propagation environment can also be represented as elements of the digital twin. If the digital twin is of the mobile communication network 200, the core network elements 223-224 can also be represented as elements of the digital twin.

[0122] The processor 205 may be configured to select a simulation model for each of the plurality of elements of the digital twin based on at least one of a minimum assessment accuracy and a maximum assessment period by determining a quantity of the plurality of elements and selecting a simulation model for each of the plurality of elements based on at least one of a minimum assessment accuracy and a maximum assessment period.

[0123] Additionally or alternatively, where one or more element types are associated with multiple simulation models of differing degrees of complexity, the processor 205 may be configured to select a simulation model for each of the multiple elements of the digital twin based on at least one of the least assessment accuracy and longest assessment period by (jointly) selecting a simulation model for each element of the one or more element types from the multiple simulation models associated with the corresponding element type based on at least one of the least assessment accuracy and longest assessment period.

[0124] 9, the processor 205 is configured to select a simulation model for each of the multiple elements of the digital twin by applying a utility function generated by the equipment 211 to the lowest assessment accuracy and / or a cost function generated by the equipment 211 to the longest assessment period. In an alternative embodiment, the processor 205 is configured to select a simulation model differently for each of the multiple elements of the digital twin.

[0125] 9, devices 201 and 211 are separate devices. In an alternative embodiment, a single device performs the functions performed by devices 201, 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.

[0126] In the embodiment shown in FIG. 9, the device 201, 211 includes one processor 205 or 215. In alternative embodiments, the device 201 and / or device 211 includes multiple processors. The processor may be a general-purpose processor, such as an Intel or AMD processor, or a special-purpose processor. The processor may include, for example, multiple cores. The processor may run, for example, a Unix-like or Windows operating system. The memory 205 and / or memory 215 may be comprised of solid-state memory, such as one or more solid-state disks (SSD) made of flash memory, or one or more hard disks.

[0127] The receivers 203, 213 and transmitters 204, 214 can communicate with each other and with other devices, for example within the RAN or CN, using one or more communication technologies (wired or wireless). The receivers and transmitters may be combined into a transceiver. The devices 201, 211 may include other elements typical of a computer server or network unit, for example a power supply.

[0128] In a typical example, an exemplary optimization problem requires the assembly of a digital network twin that includes a combination of high and low complexity elements. In this example, it becomes necessary to periodically optimize the antenna downtilt in response to spatiotemporal variations in offered traffic load.

[0129] In this typical example, the objective of adjusting the antenna down-tilt is to optimize the offered quality of service for a given load scenario, while still guaranteeing some minimum (e.g., 99.5%) coverage. It is easy to see that not only the absolute load level but also the spatial distribution of traffic will affect the optimal tilt setting. For example, if all users are located close to the base station site, the optimal down-tilt is likely to be larger than if the users are spread evenly throughout the area or if the distribution is biased towards the cell edge. Since users tend to move over time (e.g., between working and non-working hours) and their communication needs change over time (e.g., work-related email during the day vs. HD video streaming at night), it is reasonable that the optimal antenna tilt setting may also change over time.

[0130] The problem definition specifies a geographical area, including e.g. 100 base stations, the antenna downtilt per sector as configuration parameter, and the composite KPI as formulated above, i.e. optimizing the quality of service (e.g. throughput rate for the 10th user) under the condition that the coverage level is greater than e.g. 99.5%. Furthermore, the urgency is marked low and quantified as a maximum assessment time up to several tens of minutes. This maximum assessment time may be included in the problem definition or specified separately. The latter is advantageous when the maximum assessment time can change while the (rest of) the problem definition remains constant.

[0131] In this example, the digital twin required to adequately model the network in the optimization problem includes at least simulation models of 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 needed that realistically captures the antenna height, azimuth, 3D radiation pattern and transmit power, since all these factors have a significant impact on the coverage, the experienced SINR and therefore the quality of service. It would be sufficient if an extensive characterization of the propagation environment, covering in particular the path loss as a function of distance, the indoor penetration loss and the log-normal attenuation factor, was included in a correspondingly selected (low complexity) simulation model.

[0132] In this example, small-scale propagation factors such as multipath fading may be excluded since the associated small-scale variations are typically addressed by, for example, adaptive modulation and coding, transmit power control and channel-adaptive packet scheduling, rather than by adjusting antenna tilt settings. Finally, an estimate of the spatial traffic distribution valid for the future time period is used. Since a user-specific level of detail is not required (macroscopic accuracy up to the level of 40x40m pixels is typically preferred) and user mobility is negligible, a medium-complexity simulation model for the spatial traffic distribution may suffice. It should be noted that in this assembled digital twin, various factors that may be relevant for other optimization studies, such as detailed modeling of various RRM mechanisms such as packet scheduling, handover management, admission / congestion control, etc., are consciously excluded.

[0133] The simulation considers multiple antenna tilt sets. To find the optimal antenna down-tilt, an initial set is selected based on the current active configuration, and then a subsequent set is intelligently selected, for example, based on a gradient-based optimization strategy. All tuples are evaluated against the selected digital twin. With the composite KPIs in mind, the optimization strategy keeps track of the best tilt set assessed so far, and once the approach has sufficiently converged, it is selected as the optimal solution and applied to the operational communication network.

[0134] As mentioned above, the simulation model may be selected at the element type level or at the element level. For illustrative purposes, three simulation models with different levels of complexity can be identified for the propagation environment of an element type: A low-complexity simulation model of the propagation environment may include only large scale factors such as distance-based path loss, indoor penetration loss and log-normal attenuation as outlined above. This model is suitable for periodic optimization of antenna downtilt. A slightly more complex (medium complexity) simulation model may further include a small contribution of multipath fading modeled as a relatively simple time / frequency domain variation of the propagation gain around a large average value. This model can be used to optimize the packet scheduler configuration. A high-complexity simulation model of the propagation environment includes all of the above factors and further includes characterization of small-scale propagation factors per transmit / receive antenna pair or per path that may be required to optimize physical layer algorithms, SU / MU-MIMO beamforming strategies and corresponding multi-user packet (joint) scheduling algorithms.

[0135] FIG. 10 is a block diagram illustrating an exemplary data processing system capable of implementing the methods described with reference to FIGS. 1-3 and 5-6.

[0136] 10, data processing system 300 may include at least one processor 302 coupled to memory element 304 via a system bus 306. As such, 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 memory capable of performing the functions described herein.

[0137] The memory element 304 may include one or more physical memory devices, such as, for example, a local memory 308 and one or more mass storage devices 310. Local memory may refer to random access memory or other non-persistent memory devices typically used during the actual execution of the program code. The mass storage devices may be implemented as hard drives or other persistent data storage devices. The processing system 300 may also include one or more cache memories (not shown) that temporarily store at least a portion of the program code to reduce the number of times the program code is retrieved from the mass storage devices 310 during execution.

[0138] Input / output (I / O) devices, depicted as input devices 312 and output devices 314, may optionally be coupled to a data processing system. Examples of input devices include, but are not limited to, a keyboard, a pointing device such as a mouse, etc. Examples of output devices include, but are not limited to, a monitor or display, speakers, etc. The input and / or output devices may be coupled to the data processing system directly or through intervening I / O controllers.

[0139] In one embodiment, the input and output devices may be implemented as a combined input / output device (illustrated in FIG. 10 by the dashed lines surrounding input device 312 and output device 314). An example of such a hybrid device is a touch-sensitive display, also known as a "touch screen display" or simply a "touch screen." In such an embodiment, input to the device is provided by, for example, the movement of a physical object, such as a stylus or a user's finger, on or near the touch screen display.

[0140] Network adapters 316 may also be coupled to the data processing system 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 a data receiver to receive data transmitted from the systems, devices, and / or networks to data processing system 300, and a data transmitter to transmit 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.

[0141] As shown in Figure 10, the memory element 304 may store an application 318. In various embodiments, the application 318 may be stored in the local memory 308, in one or more mass storage devices 310, or separately from the local memory and the mass storage device. It should be appreciated that the data processing system 300 may further execute an operating system (not shown in Figure 10) that may facilitate the execution of the application 318. The application 318, implemented in the form of executable program code, may be executed by the data processing system 300, for example, by the processor 302. In response to the execution of the application, the data processing system 300 may be configured to perform one or more operations or method steps described herein.

[0142] Various embodiments of the invention may be implemented as a program product for use with a computer system, the program of the program product defining the functions of the embodiment (including the methods described herein). In one embodiment, the program may be included in various non-transitory computer readable storage media, and the expression "non-transitory computer readable storage media" as used herein includes all computer readable media, the only exception being a transitory propagating signal. In another embodiment, the program may be included in various transitory computer readable storage media. Exemplary computer readable storage media include, but are not limited to: (i) non-writeable storage media in which information is permanently stored (e.g., a CD-ROM disk readable by a CD-ROM drive, a ROM chip, or a read-only memory device in a computer such as any type of solid-state non-volatile semiconductor memory), and (ii) writeable storage media in which changeable information is stored (e.g., a flash memory, a floppy disk in a diskette drive or a hard disk drive, or any type of solid-state random access semiconductor memory). The computer program is operable on the processor 302 described herein.

[0143] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly dictates otherwise. Furthermore, it will be understood that the terms "comprises" and / or "comprising," when used herein, specify the presence of referenced features, integers, steps, operations, elements, and / or elements, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, elements, and / or groups thereof.

[0144] Corresponding structures, materials, acts, and equivalents of all means or steps in the following claims as well as functional elements are intended to include any structure, material, or act that performs a function in combination with elements recited in other claims as specifically recited in the claims. The description of the embodiments of the present invention has been presented for illustrative purposes, but is not intended to be exhaustive or limited to the practice in the form disclosed. Many changes and modifications will become apparent to those skilled in the art without departing from the scope of the invention. The embodiments have been chosen and described in order to best explain the principles and some practical applications of the invention, and to enable others skilled in the art to understand the invention through various embodiments with various modifications as suited to the particular uses contemplated.

Claims

1. A system (201, 210) for performing a simulation on a digital twin (1) of a telecommunications network (200), comprising at least one processor (205, 215) including: - determining a minimum assessment accuracy (66) and / or a maximum assessment period (67) for one or more performance indicators, - selecting a simulation model (13-15, 23-25, 33-35, 43-45) for each of a plurality of elements (3-8) of the digital twin (1) based on at least one of the minimum assessment accuracy and the maximum assessment period; - Assembling the selected simulation models into the digital twin (1); - running the simulation on the digital twin (1) The system is configured as follows:

2. The system (201, 210) of claim 1, wherein the at least one processor (205, 215) is configured to select the simulation models (13-15, 23-25, 33-35, 43-45) for each of the plurality of elements (3-8) of the digital twin (1) based on at least one of the lowest assessment accuracy and the longest assessment period, by determining a quantity of the plurality of elements based on at least one of the lowest assessment accuracy and the longest assessment period and selecting a simulation model for each of the plurality of elements (3-8).

3. 3. The system (201, 210) according to claim 1 or 2, wherein, for at least one of the plurality of elements, an element type (11, 21, 31, 41) of each element is associated with a plurality of simulation models (13-15, 23-25, 33-35, 43-45) having different degrees of complexity, and the at least one processor (205, 215) is configured to select a simulation model for each of the plurality of elements (3-8) of the digital twin (1) based on at least one of the lowest assessment accuracy and the longest assessment period by selecting at least one simulation model of the at least one element (3-8) from the plurality of simulation models (13-15, 23-25, 33-35, 43-45) based on at least one of the lowest assessment accuracy and the longest assessment period.

4. The system (201, 210) of any one of claims 1 to 3, wherein the at least one processor (205, 215) - running a simulation on said digital twin (1) to determine one or more network configuration settings of said telecommunications network (200); - configuring said telecommunications network (200) with one or more network configuration settings; The system is configured as follows:

5. The system (201, 210) of any one of claims 1 to 4, wherein the at least one processor (205, 215) - selecting a problem definition from a plurality of problem definitions, each of which specifies one or more configuration parameters to be optimized and corresponding one or more performance indicators; - selecting the simulation model for each of the plurality of elements (3-8) of the digital twin (1) further based on the selected problem definition; - selecting one or more configuration parameters for the simulation from the selected problem definition; The system is configured as follows:

6. The system of claim 5 , wherein the selected problem definition comprises the minimum assessment accuracy and / or the maximum assessment period.

7. The system (201, 210) of any one of claims 1 to 6, wherein the at least one processor (205, 215) Assembling multiple simulation models into different digital twin candidates (70-73); - running a number of simulations for each of the different assembled digital twin candidates (70-73) to determine a training sample; - generating a utility function (61) for the one or more performance indicators based on the training samples for a plurality of digital twin candidates (71-73) including the assembled different digital twin candidates; - generating a cost function (62) for the plurality of digital twin candidates (70-73) based on the training samples; - selecting the simulation model for each of the elements of the digital twin (1) by applying the utility function (61) to the least accurate assessment (66) and / or by applying the cost function (62) to the longest assessment period (67); The system is configured as follows:

8. 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 values ​​of the one or more performance indicators.

9. In the system (201, 210) described in any one of claims 1 to 8, the at least one processor (205, 215) is configured to select the simulation model for each of the plurality of elements (3-8) by selecting one digital twin candidate from a plurality of digital twin candidates (71-73) based on at least one of the lowest assessment accuracy (66) and the longest assessment period (67), and further based on expert knowledge.

10. A 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 plurality of elements (3-8) of the digital twin (1) based on a maximum assessment period (67) such that the simulation for the digital twin (1) provides maximized assessment accuracy without exceeding the maximum assessment period (67).

11. A 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 a plurality of elements of the digital twin (1) based on a minimum assessment accuracy (66) such that the simulation for the digital twin (1) results in a shortest assessment period while satisfying the minimum assessment accuracy (66).

12. A system (210, 211) for generating utility and cost functions (61, 62) for candidate digital twins of a telecommunications network, the system (210, 211) comprising at least one processor (205, 215): Assembling a plurality of simulation models (13-15, 23-25, 33-35, 43-45) into different candidate digital twins (70-73) of a telecommunications network (200); - running a number of simulations for each of the different assembled digital twin candidates (70-73) to determine a training sample; - generating a utility function (61) for one or more performance indicators based on the training samples for a plurality of digital twin candidates (70-73) including the assembled different digital twin candidates; - generating a cost function (62) for the plurality of digital twin candidates (70-73) based on the training samples; - storing in a memory (207, 217) information specifying the cost and utility functions (61, 62) generated for the plurality of digital twin candidates (70-73); A system configured as follows.

13. 1. A computer-implemented method of performing a simulation on a digital twin of a telecommunications network, comprising: - determining (101) a minimum assessment accuracy for one or more performance indicators and / or determining a maximum assessment period; - selecting (103) a simulation model for each of a plurality of elements of the digital twin based on at least one of the minimum assessment accuracy and the maximum assessment period; - assembling (105) the selected simulation models into the digital twin; - running (107) said simulation on said digital twin; The method includes:

14. 1. A computer-implemented method for generating utility and cost functions for candidate digital twins of a telecommunications network, comprising: - assembling (175) the simulation model into different candidate digital twins of the telecommunication network; - running a number of simulations for each of the different assembled digital twin candidates to determine a training sample (177); - generating (179) utility functions for one or more performance indicators based on the training samples for a plurality of candidate digital twins including the different assembled candidate digital twins; - generating (181) a cost function for the plurality of candidate digital twins based on the training samples; - storing in memory (183) information specifying the generated cost and utility functions for the plurality of candidate digital twins; The method includes:

15. A computer program or suite of computer programs comprising at least one software code portion, or a computer program product storing at least one software code portion, said computer program or suite of computer programs being configured to perform the method according to claim 13 or 14 when said software code portion is executed on a computer system.

Images