Method for optimizing the allocation of frequencies to cells in a mobile communication network - Patents.com

Abstract

The present invention relates to a computer-implemented method for optimizing the allocation of frequencies (f) to cells (C) of a mobile communication network (1). The cells (C) are distributed for communication in the mobile communication network (1). A set of unplanned cells (2) in the mobile communication network (1) is designated. For each unplanned cell (2), a set of frequencies (f) potentially assigned to this unplanned cell (2) is designated. A frequency interference probability (p) of selected cell pairs is calculated, each cell pair defining a relationship between the unplanned cell (2) and another cell (c) in the mobile communication network (1). Then, stress function terms are formulated, each term linking the calculated frequency interference probability (p) of the respective cell pair to a frequency relationship between the cells (2, c) of the respective cell pair. Using a quantum-inspired processor (7), an optimized allocation of frequencies (f) is determined, whereby for each unplanned cell (2), a subset of frequencies (f) is selected from the respective set of frequencies such that the stress function is minimized.

Description

[Technical field]

[0001] The present invention relates to a computer-implemented method for optimizing the allocation of frequencies to cells of a mobile communication network, the cells being distributed for communication within the mobile communication network. The present invention also relates to a quantum-inspired processor configured to perform such a method, as well as a computer program implemented to perform such a method. [Background technology]

[0002] The communication demands in terms of traffic in mobile communication networks are constantly increasing. More and more devices and applications are pushing mobile communication to new peaks. In addition, the increasing demand for digitalized and decentralized mobile work, as well as the increasing demand for streaming, are other major factors for this trend. The increasing amount of data transported through mobile communication networks, along with the demand for increased network capacity, poses significant challenges to service providers.

[0003] Mobile communication networks, particularly mobile telephone networks, e.g. 2G networks such as GSM or TETRA, include a topology of distributed "cells" for transmitting, receiving and forwarding mobile communication traffic and control data within the network via radio frequency communication. Each cell serves as a mobile communication node covering an area for the mobile communication network, thereby using one or more communication frequencies. Mobile communication participants, such as mobile phones and other cellular or wireless communication devices, connect to the respective cells for mobile communication with other participants in the communication network.

[0004] Depending on the location and characteristics of the cell's surroundings (e.g. urban or rural area), and depending on the number and density of participating devices, cells have different sets of frequencies assigned to them in order to ensure the amount and volume of communication traffic. The frequencies are used for data or speech transmission, control channels, etc. Thus, to avoid congestion in mobile communication networks and degradation of the user experience, mobile telephone providers are interested in making the best possible use of the frequency spectrum offered to them.

[0005] The capacity of a mobile communication network can be improved by densely allocating frequencies from the frequency spectrum of the mobile network operator to cells. However, this can have the drawback that the frequencies of neighboring cells violate the necessary distance requirements. For example, if the frequencies of neighboring or related cells are equal or adjacent in the frequency spectrum, they will interfere with each other. This leads to signal interference, deterioration of communication quality, or communication failure.

[0006] One possibility to solve the problem of frequency allocation in cellular communication networks involves a two-step approach. In a first step, a preliminary allocation is made with respect to conditions to be met (hard constraints), such as conditions on neighboring cells. The second step consists of an iterative brute-force local optimization in unsaturated cells. For each such cell, a ring of direct neighbouring cells is re-planned. If no satisfactory solution is found, the procedure can be repeated from the first step of ring generation with a wider ring that is the union of all cells from the neighboring ring of directly adjacent cells. This ring-based cell collection procedure can be performed recursively. However, such a conventional approach has the drawback that local optimization in unsaturated cells aggravates interference in other cells in the network, or even leads to the fact that certain cells are not or insufficiently assigned the required number of frequencies. Moreover, such a conventional approach has the drawback that the optimization techniques applied so far reach their limits rapidly. Summary of the Invention [Problem to be solved by the invention]

[0007] It is therefore an object of the present disclosure to provide an improved technique that allows an optimized allocation of requested frequencies to cells of a mobile communication network, thereby avoiding intra-cell and inter-cell interference and improving optimized frequency usage within the network. [Means for solving the problem]

[0008] This problem is solved by the method according to claim 1. Further implementations are described in the dependent claims and the following description.

[0009] The method is a computer-implemented procedure for optimizing the allocation of frequencies to cells of a mobile communication network, the cells being distributed for communication within the mobile communication network.

[0010] In the following, "unplanned cells" are referred to as mobile network cells that are defined by the coverage area of ​​the antennas that must have one or more frequencies allocated to them.

[0011] The method includes the following steps. specifying a set of unplanned cells in the mobile communication network; · for each unplanned cell, specifying a set of frequencies potentially assigned to this unplanned cell; - calculating a frequency interference probability for selected cell pairs, each cell pair defining a relationship between an unplanned cell and another cell in the mobile communication network; formulating stress function terms, each term relating a calculated frequency interference probability of a respective cell pair to a frequency relationship between the cells of the respective cell pair; determining, for each unplanned cell, an optimized allocation of frequencies by selecting a subset of frequencies from the respective set of frequencies such that a stress function is minimized by using a quantum inspired processor;

[0012] The method robustly addresses the problem of allocating frequencies to unplanned (or to be planned) cells in a communication network, such that for most (or ideally all) unplanned cells, a required subset of selected frequencies can be allocated according to the frequency demand of the cell. This allocation satisfies certain proximity conditions between adjacent cells. Optionally, the overall frequency interference between cells in the mobile communication network is minimized.

[0013] By applying this method, for each given unplanned cell, an optimal allocation of a subset of frequencies from the set of frequencies potentially assigned to this cell can be selected, which is chosen automatically, such that intra-cell and inter-cell frequency interference can be minimized, while still improving dense frequency usage across the available frequency spectrum in the network.

[0014] Each subset of frequencies reflects a certain required number of frequencies per cell, or a certain required frequency of the frequency spectrum per cell. In other words, this means that for a certain required number of frequencies per cell, a subset of that size (=concentration) of the frequency spectrum of that cell must be determined. Here, frequencies are defined as discrete frequency bands or channels, for example in a GSM network. Here, a cell is defined as a certain area coverage of mobile communications, for example by one or more antennas for wireless communication. Depending on the transmission power characteristics and installation of the antennas, as well as the topology of their surroundings, the area coverage of the cell covered by the antennas may vary.

[0015] An exemplary mobile communications GSM network comprises thousands of GSM cells and tens or hundreds of frequency bands or channels, with a small number of channels (approximately <10, eg 2-6) used per cell.

[0016] A frequency relationship in this context describes a frequency-dependent expression that defines a (potential) relationship between one or more frequencies of one cell of each defined cell pair and one or more frequencies of the other cell. For example, a frequency relationship describes certain combinations of selected frequencies for both cells of a cell pair that may result in interference between the cells of the cell pair with respective calculated frequency interference probabilities.

[0017] A cell pair in this context describes the relationship of each cell from the set of unplanned cells to another cell in the mobile communication network. Said other cell may be a cell from the set of unplanned cells to which a frequency is assigned. Alternatively, said other cell may be a cell in the mobile communication network that is not in the selected set of unplanned cells. It may for example be a previously planned cell with a pre-assigned frequency or a cell with a fixed frequency that may for example result from a previous assignment.

[0018] For example, a frequency assignment FA can be described as a mapping from a set of cells to a power set (a set of subsets) of frequencies. The interference probability per cell pair can be defined as a real-valued function of FA×FA that assigns a probability of disturbance to each cell pair of “frequency-assigned cells”. For a frequency assignment as a whole, an interference probability can be defined (e.g., as the sum of pairwise interference probabilities for all off-diagonal pairs of FA×FA). This can be interpreted as a degree of disturbance in the whole network, rather than a probability per se (e.g., the value can be >1). The method can be implemented to search for the one with the smallest interference probability in the set of all possible frequency assignments FA. Further constraints may be optionally considered, as will be further explained below.

[0019] A stress function (cost function) for the optimization problem described above can be constructed by calculating the frequency interference probability of selected cell pairs and formulating stress function terms, whereby each term relates the calculated frequency interference probability of a respective cell pair to the frequency relationship between the cells of said respective cell pair. Such a stress function penalizes frequency relationships that result in significant interference between the cells. The stress function, and hence the underlying optimization problem, can be formulated as a quadratic unconstrained binary polynomial.

[0020] In this way, an optimized allocation of frequencies to unplanned cells is determined by selecting, for each unplanned cell, a subset of frequencies from the respective set of frequencies such that a stress function is minimized. The minimum of the stress function is preferably a global minimum, but may also be a local minimum.

[0021] Thus, the method has the technical effect and advantage of an optimized coverage of cells with the required number of frequencies for optimal frequency utilization per cell depending on the location of the cell and the characteristics of the surroundings of the cell, as well as depending on the required demand derived from the expected number and density of communication participants. Furthermore, interference minimization can be reached in an optimal manner. Thus, the user experience can be improved and degradation of communication quality or communication failure can be avoided. Furthermore, the cost of installation and maintenance of cells can be reduced, since the number of cells required can be reduced due to the optimized frequency allocation. Furthermore, the method has the effect and advantage that the solution of the optimized allocation of frequencies to unplanned cells is determined / calculated very quickly, thereby enabling the possibility of very fast replanning / reallocation of frequencies to the respective cells.

[0022] The underlying optimization problem can be very complex. This is not only because the frequency of one cell can affect the frequency of other cells. It is also very complex because of the many practical constraints that must be taken into account. As more constraints are implemented, such problems become more complex and difficult to solve. This is problematic or difficult when a solution is required quickly, for example, when a frequency reassignment to certain cells is required as a reaction to an unexpected network failure or during maintenance. However, even when further practical constraints such as neighborhood relationships between cells or dense agglomerations of cells in so-called "sectors", for example several antennas on rooftops in an urban area or city, are taken into account, such calculations are difficult to solve. The method described herein advantageously shows its strengths compared to conventional approaches the more complex the underlying problem becomes. In other words, for complex optimization problems taking into account practical constraints such as those described above, the method described herein has significant advantages over conventional techniques.

[0023] One advantage of the procedure described herein is the automatic solution of the optimization problem without manual adjustments such as the number of rings in the conventional approach as described above. The execution time is shorter and the quality of the solution is better compared to state-of-the-art procedures, since more required frequencies can be reliably assigned to more cells in the network. The procedure described is a hybrid approach between a pre-processing step (forming a set of unplanned cells and / or forming a respective set of frequencies to be potentially assigned to each unplanned cell) and a solution step by a quantum-inspired processor.

[0024] The method described herein utilizes a quantum-inspired approach. The calculation of the optimized solution of the stress function for determining the optimized allocation of frequencies to all cells of the set of unplanned cells is performed by a so-called quantum-inspired processor. As a quantum-inspired processor in the context of the present disclosure, a processor that solves the so-called "Ising model" or equivalent quadratic unconstrained binary problem is defined. For example, this is a processor configured to solve the optimization problem by quantum annealing or quantum annealing emulation. Such a processor is, for example, based on conventional hardware technology, for example based on complementary metal-oxide semiconductor (CMOS) technology. An example of such a quantum-inspired processor is the Fujitsu Digital Annealer. Alternatively, any other quantum processor can be used for the method described herein, and in the future, also technologies based on actual qubit technology can be used. Further examples of such quantum-inspired processors are the quantum annealers from DWave (e.g., 5000Q), but also quantum gate computers (IBM, Rigetti, OpenSuperQ, IonQ or Honeywell) and their future successors, or alternative QC designs utilizing quantum optimization algorithms such as the Quantum Approximate Optimization Algorithm (QAOA) or the Variational Quantum Eigensolver (VQE).

[0025] In other words, a quantum-inspired processor as defined herein is a processor that implements the concept of minimization of so-called quadratic unconstrained binary optimization (QUBO) functions, either on specialized processors based on classical techniques, on quantum gate computers, or on quantum annealers.

[0026] In at least one implementation, the method further includes the steps of: - specifying frequency variables, each frequency variable being associated with one cell in the mobile communication network and one frequency for this cell; formulating said frequency relationships in terms of said stress function as combinations of selected frequency variables; · Calculating the terms of the stress function by using a quantum inspired processor and setting the frequency variables such that the stress function is minimized.

[0027] The frequency for the cell to which each frequency variable is associated is, for example, one frequency from a set of frequencies potentially assigned to the cell, if the cell is a cell that has not yet been planned. Alternatively, such a frequency is, for example, one frequency already assigned to the cell, if the cell has already been planned. In the latter case, the value of the frequency variable is known a priori depending on whether the frequency is assigned to the cell (value 1) or not (value 0). Such fixed variables do not need to be optimized by the quantum-inspired processor, but may lead to constraints for the unplanned cells and thus may be part of some linear terms of the stress function. The actual number of frequency variables to be optimized is only those coming from the unplanned cells. For example, for n cells with m frequencies per cell, at least n×m frequency variables may be specified.

[0028] In this way, for each unplanned cell, the optimal allocation of frequencies can be calculated individually, such that the stress function is minimized by setting the frequency variables respectively. This provides an elegant implementation of a very flexible and variable frequency allocation. Thus, different subsets of frequencies for different cells can be selected to avoid disturbing interference between cells or a significant increase in interference within the network. Furthermore, dense frequency usage can be achieved in an optimized manner across cells distributed throughout the network.

[0029] The formulation of the frequency relationship in terms of the stress function as a combination of selected frequency variables allows the calculation of an optimal solution (minimum) of the stress function for all unplanned cells in a flexible manner taking into account various frequency scenarios that may result in different disturbance interference behavior between cells. In this way, an optimized selection of different subsets of frequencies for all unplanned cells can be achieved so as to minimize the cost function of the above mentioned optimization problem. Thus, the influence of the selected frequency of one cell on other (potentially assigned) frequencies of other cells can be mitigated. This allows a very high degree of freedom in frequency allocation, but is still very complex to solve. The optimized selection of the respective subsets of frequencies for all unplanned cells is performed by a quantum-inspired processor as described above.

[0030] In at least one implementation of the method, only calculated frequency interference probabilities below a certain threshold are considered in the stress function term. Thus, the optimization considers only interference probabilities that do not exceed a certain limit. This prevents certain frequencies from being assigned to certain cells that may result in unacceptable interference between these cells. For example, interference probabilities above a threshold are avoided by a penalty term that is forced to zero for the optimal solution by a large prefactor.

[0031] In at least one implementation, the method is performed taking into account frequency demand conditions such that each subset of frequencies allocated to unplanned cells has a specified number of frequencies. One frequency is defined as a control channel frequency and the other frequencies are defined as traffic channel frequencies. In a different implementation, the control channel frequency can be just one or at least one frequency for each cell. The number of frequencies allocated to each cell is, for example, preselected or predetermined taking into account, for example, the topology of the mobile communication network, the location and / or cellular characteristics of each cell (which depends on the power transmission performance and / or topology around the cell), as well as the number and density of participants who wish to dial into said cell for mobile communication.

[0032] According to the frequency demand conditions, each cell must be assigned one control channel frequency. One control channel frequency is required for each cell, which serves the purpose of providing a control channel for the cell, via which signaling and control information is exchanged between the cell and other cells or between the cell and different network elements, terminals or participants. If the mobile communication network is implemented according to the GSM standard, such a control channel frequency is, for example, a so-called broadcast channel (BCCH) frequency according to the GSM standard. Furthermore, other frequencies that are assigned to the cell according to the frequency demand conditions are defined as traffic channel frequencies, which serve the purpose of providing one or more traffic channels for the cell, through which communication traffic, such as data or voice, is transmitted and exchanged between the cell and other cells or between the cell and different network elements, terminals or participants. If the mobile communication network is implemented according to the GSM standard, such traffic channel frequencies are, for example, so-called transmission channel (TCH) frequencies according to the GSM standard.

[0033] With respect to controlling communication between cells in a mobile communication network, control channels have a more fundamental role than traffic channels. In some implementations, control channels are configured to transmit communication traffic in addition to signaling and control information. In other implementations, control channels are configured to transmit only signaling and control information.

[0034] In at least one implementation, the method is performed taking into account a frequency combination distance condition, in which for each unplanned cell, a forbidden frequency relationship is defined such that the frequencies of the subset of frequencies allocated to this cell have a certain frequency relationship with a certain channel distance, for example, the channel distance being equal to or greater than 2. Thus, the frequency combination distance condition has the effect that intra-cell frequency interference can be minimized or completely avoided, since the frequencies used by each cell must satisfy a certain channel distance relative to each other.

[0035] In at least one implementation, the method is performed taking into account cell proximity conditions, whereby for cells having a determined (handover) proximity relationship, prohibited frequency relationships between these cells are defined. The proximity relationship in particular concerns neighboring cells (e.g. a selected cell pair or a plurality of cells) that provide an overlap in their area coverage. This is important in view of so-called handovers between neighboring cells. If a participant moves away from the antenna (center) of one cell during ongoing use of the mobile communication network, its transmission level decreases. As soon as the participant reaches the area of ​​the next neighboring cell, the participant has to be handed over from the first cell to the second cell in order to continue using the mobile communication network. In such a scenario, if certain frequency relationships were used, the inter-cell frequency interference between such neighboring cells would increase beyond a critical limit. In such a case, the necessary calculations for handover may be disturbed by interference and may be erroneous or fail, with the risk that the handover will fail and the participant will face a communication interruption in the mobile communication network. Thus, to avoid significant frequency interference between these cells, and in particular to prevent potential drawbacks of handover failure, the cell proximity condition defines forbidden frequency relationships between neighboring cells.

[0036] In at least one implementation, the method is performed taking into account a cell-sector condition, where for cells having a determined sector relationship, a prohibited frequency relationship between these cells is defined. The cell-sector relationship relates to cells in a given dense installation area or location, for example, where multiple antennas of different cells are installed close to each other, for example, on a rooftop, in a city center, or in a dense urban area. Here, inter-cell interference may be significant due to the dense installation and overlap of different cells. Therefore, in a sector, the prohibited frequency relationship between cells is very important to prevent communication failure. The cell-sector condition may be more restrictive than the cell neighborhood condition, meaning that the cell-sector condition defines stricter rules for the prohibited frequency relationship than the cell neighborhood condition.

[0037] In at least one implementation, the method is performed taking into account frequency interference probability conditions, thereby defining prohibited frequency relationships between cells of cell pairs in a mobile communication network where the frequency interference probability is above a predefined threshold. In this way, frequency interference between cells can be kept within acceptable limits throughout the network, preventing adverse effects that could lead to communication failure in the worst case. Even if the overall allocation of frequencies after all unplanned cells have been planned and assigned a subset of their frequencies results in an increased interference probability between some cells of the network or within certain areas of the network, the local and global interference probability can still be kept within acceptable limits.

[0038] In at least one implementation of the method, frequency relationships of the type described above distinguish between co-channel and adjacent channel frequency relationships. Co-channel frequency relationships consider frequencies of the same channel. Adjacent channel frequency relationships consider frequencies of adjacent channels. Adjacent channels mean directly adjacent frequency channels. For example, channels f3 and f4 are adjacent channels, but channels f3 and f5 are not adjacent channels.

[0039] In this implementation, co-channel and adjacent channel frequencies are considered to contribute to interference within one cell or between each cell pair. In general, co-channel frequencies have a high impact on interference because the frequencies use the same channel. When adjacent channel frequencies are used, the impact on interference is not as high as that of co-channel frequencies. However, interference can be a problem even between adjacent channel frequencies because those frequencies have a small frequency distance (narrow frequency band spacing). For example, if two or more of the frequencies assigned to one cell are co-channel or adjacent channel frequencies, high interference will be a problem, greatly affecting the communication quality of this cell or resulting in communication failure. Similarly, if two cells having a certain relationship, for example, a neighborhood relationship or sector relationship as described above, use two or more co-channel or adjacent channel frequencies, interference may increase significantly with respect to these cells, which may also affect the communication quality or result in communication failure. Furthermore, if two cells use two or more co-channel or adjacent channel frequencies and the calculated frequency interference probability of at least one of these cells with the other exceeds a predetermined threshold for one or more of these frequencies, the interference may increase beyond acceptable limits and communication quality may deteriorate.

[0040] Thus, the assignment of co-channel or adjacent channel frequencies to or within cell pairs that have a certain cell relationship or have a certain frequency probability that exceeds an acceptable limit should be avoided, otherwise co-channel or adjacent channel frequency assignments may be allowed for cell pairs if the interference is within acceptable limits.

[0041] The distinction between co-channel and adjacent channel frequency relationships allows the definition of different degrees of freedom in frequency allocation for different intra-cell and inter-cell frequency scenarios, depending on how much interference may or may not be tolerable. For example, different prohibited frequency relationships may be defined, as described above, for different frequency combination distance conditions, cell proximity conditions, cell sector conditions, or frequency interference probability conditions, each of which includes a different combination of co-channel and adjacent channel frequency relationships.

[0042] In at least one implementation of the method, the frequency relationship distinguishes between a control channel frequency and a traffic channel frequency. The control channel frequency and the traffic channel frequency may be defined as described above. Depending on the frequency allocation within or between cells and depending on the relationship between the respective cell pairs, the control channel frequency and the traffic channel frequency may each have a different importance with respect to the adverse effects caused by frequency interference. For example, if the control channel has a more fundamental role than the traffic channel, the respective frequency relationships may be defined such that an optimized frequency allocation keeps the interference effect on the control channel frequency within a limit smaller than the tolerable interference limit on the traffic channel frequency.

[0043] Thus, the distinction between control channel frequencies and traffic channel frequencies in each frequency relationship allows the definition of different degrees of freedom in frequency allocation for different intra-cell and inter-cell frequency scenarios depending on the different importance of the control channels and traffic channels and how much interference for both may be tolerable. For example, for different frequency combination distance conditions, cell proximity conditions, cell sector conditions, or frequency interference probability conditions, different prohibited frequency relationships may be defined, each involving different combinations of control channel frequencies and traffic channel frequencies, as described above.

[0044] In at least one implementation that takes into account at least one of a frequency demand condition, a frequency combination distance condition, a cell proximity condition, a cell sector condition, or a frequency interference probability condition as described above, the method further includes the following steps. - specifying frequency variables, each frequency variable being associated with one cell in the mobile communication network and one frequency for this cell; formulating a frequency relationship for each condition in terms of selected frequency variables; Calculating a frequency relationship for each condition by using a quantum inspired processor and setting the frequency variables such that each frequency relationship is zero.

[0045] In this way, each condition can be flexibly modeled and optimized by setting the frequency variables, respectively, using a quantum-inspired processor, such that each condition is satisfied if possible. This provides an elegant implementation of each condition as a (hard) constraint to the optimization problem formulated in the stress function. By taking such constraints into account in the formulation or calculation of the stress function, solutions of the optimization problem that violate the above conditions can be penalized. This makes it possible to find an appropriate optimized solution, taking into account the practical constraints of the real network conditions of the communication network. The frequency variables used in formulating the frequency relations of each condition can be partially or completely the frequency variables used in formulating the terms of the stress function as described above.

[0046] In at least one implementation of the method, the calculated frequency interference probability for each cell pair distinguishes between a co-channel frequency interference probability and an adjacent channel frequency interference probability, where the co-channel frequency interference probability is calculated for frequencies of the same channel between the cells of the respective cell pair, and the adjacent channel frequency interference probability is calculated for frequencies of adjacent channels between the cells of the respective cell pair. In this way, different interference probabilities can be calculated and taken into account depending on different interference characteristics of the co-channel or adjacent channel frequencies, as described above in the context of the co-channel or adjacent channel frequency relationship.

[0047] In at least one implementation of the method, the stress function is formulated as a quadratic unconstrained binary optimization (QUBO) function. This QUBO function serves as an "input" for a quantum-inspired processor that solves this optimization problem for an optimized allocation of frequencies according to the method described above. Generally speaking, a QUBO is a quadratic polynomial in binary variables that are represented in the quantum-inspired processor as bits or quantum bits (hereafter qubits, Qbits). In the context of the optimization problem of the present disclosure, the QUBO function represents the sum of the potential contributions of the calculated frequency interference probability of each cell pair as a function of the different qubits. Each qubit represents the selection of one frequency of one cell and can achieve the value "0" or the value "1" (or both with some probability). To solve the optimization problem, the quantum-inspired processor runs through different settings of the different qubits to find such a solution that minimizes the optimization problem. In this way, the QUBO representation of the optimization problem has elegant properties with respect to the quantum-inspired computation applied here. For example, the frequency variables described above are formulated in the form of such qubits.

[0048] In at least one implementation of the method, the stress function and one or more of the conditions described above are combined into a global QUBO function. Thus, the global QUBO function can take into account one or more of the constraints described above and the optimization target. In this regard, one or more of the above-mentioned constraints can be weighted in the QUBO function as soft constraints. In this regard, a soft constraint is a constraint that is desirable but does not need to be satisfied if the cost function achieves a substantially better value. This has the advantage that the QUBO function can be fine-tuned to some extent, depending on whether the optimization problem is focused on optimizing frequency allocation to cells throughout the network with respect to interference probability, or on the satisfaction of one or more of the above-mentioned (soft) constraints.

[0049] In at least one implementation, the method further includes the following steps. designating a subset of unplanned cells from said set of unplanned cells; performing said method for a subset of said unplanned cells; updating the cells in said subset with their determined optimized frequency allocation to planned cells; · performing said method iteratively on the remaining unplanned cells of the set of unplanned cells until all unplanned cells of the set of unplanned cells have been processed.

[0050] In this way, a kind of decomposition strategy can be followed. This is advantageous or even necessary to process the described method despite the limited hardware performance of the quantum inspired processor for solving the optimization problem. As mentioned above, let us assume a large number of necessary frequency variables to formulate the optimization problem and / or one or more of said additional conditions / constraints. Today, the possibilities of current quantum inspired processors are still limited. Thus, a very complex optimization problem must be decomposed into several partial solutions that can be processed iteratively in order to find the optimal solution. At each iteration step, one partial solution is found by the quantum inspired processor.

[0051] For example, hardware limitations appear under consideration of the QUBO formulation of the optimization problem, as discussed above, and the number of bit variables (qubits) required to formulate the optimization problem depends, for example, on the number of cells and the number of frequencies (potentially assigned or already planned) to each cell. Thus, for n unplanned cells with m frequencies per cell, at least n×m qubits must be specified. Qubits for already planned cells can be set constant. If, for each frequency, furthermore, control channel and traffic channel characteristics are differentiated in their respective frequency relationships, the number of required qubits further increases to 2×n×m, since n×m qubits for control channel frequency assignment and n×m qubits for traffic channel frequency assignment must be specified.

[0052] Considering for example a GSM communication network that may contain thousands of GSM cells to be planned and tens or hundreds of frequency bands or channels, the number of qubits required can quickly reach the order of magnitude of hundreds of thousands. Considering further that a typical quantum-inspired processor today is capable of solving optimization problems on the order of 10,000 bit variables (qubits), the overall optimization problem must be solved iteratively with the aid of problem decomposition by the above mentioned means.

[0053] According to an exemplary strategy for specifying / constructing a subset of unplanned frequencies, the following steps are performed: In a first step, the candidates of unplanned cells are sorted / prioritized in descending order according to the number of certain constraints or cell relationships with already planned cells. For example, first, one or more unplanned cells are considered that have the most cell relationships with already planned cells or whose frequency allocation procedures are most affected in terms of potential frequency interference by the constraints of already planned cells. This has the advantage that first, one or more unplanned cells are considered whose planning (frequency allocation) is most constrained by interference characteristics with already planned cells in the network. These unplanned cells are then included first in the subset of unplanned cells.

[0054] In a second step, further candidates for unplanned cells are selected according to certain constraints or cell relationships with respect to the cells already included in the subset of unplanned cells. This has the advantage that in the second step, further unplanned cells are taken into account whose planning (frequency allocation) may have the most impact in terms of interference characteristics with respect to the cells already present in the subset of unplanned cells. These unplanned cells are then also included in the subset of unplanned cells.

[0055] The first step and / or the second step are performed iteratively until an upper limit of frequency variables (qubits) that can be calculated at one time by the quantum-inspired processor is reached. Then, the method according to the implementation described above is performed for the constructed subset of unplanned cells. After the frequency allocations are calculated for the subset of cells, they are updated to the planned cells with their determined optimized frequency allocations. The method is then performed iteratively for the remaining unplanned cells of the set of unplanned cells until all unplanned cells have been processed.

[0056] The above-mentioned problem is also solved by a quantum-inspired processor as set forth in the appended claims. The quantum-inspired processor is configured to perform one or more steps of the above-mentioned method. According to an exemplary implementation, the quantum-inspired processor is a digital annealing processing unit. This unit may be specially configured to perform quantum annealing or quantum annealing emulation as described above. The quantum-inspired processor may be of any type as described above.

[0057] Furthermore, the above problem is also solved by a computer program comprising instructions which, when executed by one or more processors, cause each of the one or more processors to perform one or more steps of the above method. At least one of these processors is, for example, a quantum-inspired processor as described above. Other processors may be configured to process preparatory or iterative steps of or for the above method by executing the computer program. The computer program may be stored in a computer-readable storage medium.

[0058] Furthermore, the above-mentioned problem is also solved by a workplace for a network planner, configured for validating an optimized allocation of frequencies determined by the above-mentioned method. Such a workplace has, for example, a validation means configured for (automated or semi-automated) validation of the optimized allocation of frequencies determined by the above-mentioned method. This helps the network planner to validate the optimization results found by the above-mentioned method. The validation means can be implemented in software and / or hardware. For example, the workplace can communicate with or be connected to a system comprising a quantum-inspired processor that executes the above-mentioned method. The results can then be carried over during the workplace.

[0059] Furthermore, the above problem is also solved by an interface device comprising one or more interfaces to cells of a mobile communication network, the cells being distributed for communication within the mobile communication network, the interface device being configured to automatically deploy an optimized allocation of frequencies determined by the above method to the cells of the mobile communication network. In this way, the optimized allocation of frequencies determined by the above method can be deployed (automatically or semi-automatically) to a number of cells of the respective mobile communication network. For example, the interface device can communicate with or be connected to the above workplace or a system comprising a quantum-inspired processor executing the above method. The result can then be carried over into the interface device.

[0060] Furthermore, as a preparatory measure for one or more of the above described steps of the computer implemented procedure, an interface can be implemented or used to read parameters from the communication network prior to each optimization and input such parameters to the described computer implemented optimization procedure. In this way, a closed loop continuous optimization is possible. The parameters include, for example, the network configuration, the neighboring information or relations of each cell of the network, the available frequencies in the network, and the expected frequency demand of each cell.

[0061] Any aspect, feature, advantage, and measure described alone or in combination with one another in the context of the methods described above may also apply to, or find analogous expression in, the aspects, features, advantages, and measures described alone or in combination with one another in the context of the quantum-inspired processors, computer programs, workplaces, and interfaces described above, and vice versa. [Brief description of the drawings]

[0062] The invention is further explained below with the aid of several figures and with regard to several implementations. [Figure 1] 1 illustrates an exemplary configuration of a communication network having multiple cells. [Diagram 2] 2 shows an exemplary frequency allocation for one cell of the network according to FIG. 1; [Diagram 3] 1A shows an exemplary schematic diagram of cell proximity conditions between an exemplary pair of cells of a network according to FIG. 1; and FIG. 1B shows an exemplary schematic diagram of cell sector conditions between exemplary cells of a network according to FIG. [Figure 4] 2 shows an exemplary schematic diagram of frequency interference probability conditions between an exemplary pair of cells of the network according to FIG. 1; [Diagram 5] 1 illustrates an exemplary mathematical formulation of a partial optimization problem. [Figure 6] 1 illustrates an exemplary mathematical formulation of a partial optimization problem. [Figure 7] 1 illustrates an exemplary mathematical formulation of a partial optimization problem. [Figure 8] 1 illustrates an exemplary mathematical formulation of a partial optimization problem. [Figure 9] 1 illustrates an exemplary mathematical formulation of a partial optimization problem. [Figure 10] 1 illustrates an exemplary mathematical formulation of a partial optimization problem. [Figure 11] 1 illustrates an exemplary mathematical formulation of a partial optimization problem. [Figure 12] 1 shows an exemplary schematic diagram of a frequency allocation algorithm. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0063] FIG. 1 shows an exemplary configuration of a part of a communication network 1 having a plurality of cells. For example, the communication network 1 is a GSM network and includes about 50,000 GSM cells. In FIG. 1, cells c1 to c5 are illustrated. Each of the cells c1 to c5 has an antenna a1 to a5, respectively. Thus, each of the cells c1 to c5 spans a coverage area of ​​the mobile wireless communication by the respective antenna a1 to a5. Depending on the transmission power characteristics and installation of the antennas a1 to a5 and the surrounding topology, the area coverage of the cells c1 to c5 spanned by the antennas a1 to a5 may vary.

[0064] In the exemplary configuration of FIG. 1, each of the cells c2-c5 provides a communication node for participants 3 (mobile phones and other cellular or wireless communication devices) that can connect to the respective cell through wireless communication through certain frequencies. Each cell c2-c5 covers a certain area for the mobile communication network. The frequencies are discrete frequency bands or channels. For example, the communication network 1 comprises 100 frequency channels, with approximately 3-5 channels used per cell. If hopping is used, the channels used per cell can be, for example, 1-12. In the exemplary configuration of FIG. 1, the cells c2-c5 each use certain frequencies that are assigned to the respective cell. The frequencies are assigned to the cells c2-c5 in such a way that interference (signal disturbance) between the cells c2-c5 (see the arrows between the cells c2-c5 in FIG. 1) can be avoided or at least kept as low as possible.

[0065] In contrast to cells c2-c5, cell c1 in Fig. 1 does not have a frequency already assigned to it. This means that cell c1 is an unplanned cell 2, and the other cells c2-c5 are respectively planned cells 6. However, cell c1 should be assigned some frequency so that interference (signal disturbance) between cell c1 and the other cells c2-c5 can be avoided or at least kept as low as possible, so cell c1 also becomes a planned cell.

[0066] In this regard, a frequency allocation procedure is performed in which an automated computer-implemented frequency allocation is performed to the unplanned cells 2 in the communication network 1. The optimization goal of such frequency allocation is to minimize frequency interference as much as possible between the respective cells of the communication network 1, while still allocating the required number of frequency channels to as many of the unplanned cells as possible.

[0067] The number of frequency channels required for each cell depends on the number of mobile communication participants 3 that dial into the network 1 and want to use the communication topology of the communication network 1. As shown in the exemplary configuration according to FIG. 1, cells c1 and c5 are associated with a minimum number of participants 3. For example, cells c1 and c5 are located in rural areas without dense communication demand. Cell c4 requires more participants 3 than cells c1 and c5. For example, cell c4 is located in a town with low demand for communication. In contrast, cells c2 and c3 each have a maximum number of participants 3 that want to use the communication network 1. Cells c2 and c3 are, for example, installed in a city with dense urban infrastructure. Thus, in the exemplary configuration of FIG. 1, cells c1 and c5 require a minimum number of frequency channels and cells c2 and c3 require a maximum number of frequency channels.

[0068] Frequencies are allocated to this unplanned cell 2 such that, for the cell c1, it causes as little frequency interference as possible to other cells in the communication network 1, but still obtains the necessary number of channels to serve the participants 3 requesting communication in the cell c1. In other words, the optimization problem to be solved for the cell c1 consists in determining an optimized frequency allocation to the cell c1 (and potentially to the other cells c2-c5, if frequency reallocation is considered for said cells) in the communication network 1. This means that for each cell c1 (and potentially other cells), a subset of frequencies is selected from the respective set of potential frequencies such that a mathematically formulated quadratic stress function (core optimization problem) is minimized. This serves the objective of selecting a respective subset of frequencies that meets the necessary frequency demand of each unplanned cell 2 in the network 1 depending on the number of participants 3 per cell, and has such a technical effect, combined with the effect that intra-cell and / or inter-cell frequency interference can be minimized, and nevertheless, a dense frequency usage over the entire available frequency spectrum in the network 1 can be improved. This has the effect that the user experience can be improved and degradation of communication quality or communication failures can be avoided, thereby still allowing as many participants as possible to participate in and use the communication network 1. The subset of frequencies allocated to each unplanned cell can be selected from a predefined set of frequencies that can potentially be allocated to the unplanned cell 2, exemplarily cell c1 in Fig. 1 (and potentially other cells).

[0069] In order to achieve the above advantageous effects, a computer-implemented algorithmic method is implemented for optimizing frequency allocation to unplanned cells 2 in a communication network 1, as will be further described below. In addition to such a core optimization problem, additional sub-optimization problems are considered, which represent conditions (constraints) to be satisfied in the optimization process. These constraints take into account that there are certain intra-cell and inter-cell frequency relationships as well as certain cell relationships that affect the allocation of frequencies to each cell in terms of optimized minimization of frequency interference and fulfillment of the required frequency demand per unplanned cell 2. Such conditions and constraints are described below.

[0070] 2 shows an exemplary configuration of a cell c3 to which a given frequency is assigned, which thereby meets certain conditions that are applicable to other cells in the communication network 1 (especially with respect to the unplanned cells 2 to which frequencies are to be assigned) and which must also be met by those cells.

[0071] FIG. 2 shows the distinction between two basic types of frequencies used within the cells of the communication network 1. These two types of frequencies distinguish between so-called Broadcast Channel (BCCH) frequencies and Traffic Channel (TCH) frequencies. In the exemplary configuration described here, one BCCH frequency must be assigned to each cell. A BCCH frequency is required for each cell and serves the purpose of providing a control channel for the cell, via which signaling and control information is exchanged between that cell and other cells or between that cell and different network elements, terminals or participants. A TCH frequency serves the purpose of providing one or more traffic channels for each cell, through which communication traffic, such as data or voice, is transmitted and exchanged between that cell and other cells or between that cell and different network elements, terminals or participants. The BCCH frequency can be configured to transmit signaling and control information as well as a part of the communication traffic in the communication network. Alternatively, the BCCH frequency is configured to transmit only signaling and control information. For example, the signal transmissions of the different participants 3 in the network 1 (see Figure 1) are arranged as time multiplexed transmissions, and each participant using a respective frequency in a respective cell is assigned a time slot (e.g. within milliseconds) during which the signaling assigned to this participant is transmitted.

[0072] As exemplarily shown for cell c3 in FIG. 2, each cell in the communication network 1 must meet its own specific frequency demand. As mentioned above, such frequency demand depends on the number of participants 3 dialing into the communication network 1, and may also depend on the topology of the mobile communication network 1, the location and / or cellular characteristics of each cell (which depends on the power transmission, performance and / or topology around the cell). The frequency demand defines the number of frequencies to be assigned to each cell. This number may be preselected or predetermined depending on the above-mentioned parameters. As exemplarily shown in FIG. 2, cell c3 has a frequency demand of 4, which means that four frequencies are assigned to cell c3. As explained, one frequency is a BCCH frequency. In the example of FIG. 2, frequency f18 is assigned as the BCCH frequency of cell c3. Also, the rest of the assigned frequencies, here three frequencies, are assigned as TCH frequencies. In the example of FIG. 2, frequencies f1, f3 and f99 are assigned to cell c3 as TCH frequencies.

[0073] As further shown in FIG. 2 , depending on whether the characteristic of the frequency is a BCCH frequency or a TCH frequency, a certain frequency variable x is associated with one cell c in the mobile communication network 1 and one frequency f for this cell c. cf is defined, where the frequency variable x cf takes the value "0" when the respective frequency f is not assigned to the respective cell c, and takes the value "1" when the respective frequency f is assigned to the respective cell c. As shown in FIG. 2 as an example, the frequency f18 is assigned to the cell c3 as the BCCH frequency, so that the frequency variable x c3f18 BCCH = 1. As mentioned above, frequencies f1, f3, and f99 are assigned to cell c3 as TCH frequencies, so the other three frequency variables are also x c3f1 TCH =1, x c3f3 TCH =1, x c3f99 TCH= 1. All other frequency variables x for cell c3 are set c3f is 0. Such a formulation using frequency variables serves the purpose of the constraints mentioned above and the mathematical formulation of the core optimization problem mentioned above, which is further explained below.

[0074] In addition to the frequency demand condition, the cell c3 exemplarily shown in FIG. 2 must also satisfy the so-called frequency combination distance condition. This means that for each cell c, in particular for each unplanned cell 2, a forbidden frequency relationship is defined such that the frequencies f assigned to each cell have a certain channel distance. In a configuration as exemplarily shown with respect to FIG. 2, such a channel distance may, for example, have a value ≧2. This means that each frequency f assigned to the respective cell must have a frequency distance ≧2 to other frequencies f of the cell. This requirement states that so-called co-channel and adjacent channel frequencies are forbidden for each cell. Co-channel frequencies relate to frequencies f of the same channel, and adjacent channel frequencies relate to frequencies f of adjacent channels. This serves the purpose of avoiding intra-cell interference between the frequencies assigned to each cell. As exemplarily shown for cell c3 in FIG. 2, the frequencies f1, f3, f18, and f99 all have a distance of ≧2, so that the frequency combination distance condition is also satisfied.

[0075] FIG. 3a exemplarily shows further conditions that must be satisfied as additional constraints. FIG. 3a shows so-called cell neighborhood conditions, which means that for cells c having a determined neighborhood relationship 4, a forbidden frequency relationship between these cells c is defined. As exemplarily shown in FIG. 3a, cells c1 and cN are considered to be in neighborhood 4. Such neighborhood relationship 4 thus describes here neighboring cells that provide an overlap in their area coverage. For cells c in the neighborhood 4, certain combinations of frequencies are forbidden to be assigned to the respective cell pair. As exemplarily shown in FIG. 3a, the BCCH frequency of cell c1 must not be a co-channel (co) or adjacent channel (adj) frequency of the BCCH frequency of cell cN. Furthermore, the BCCH frequency of cell c1 must not be a co-channel or adjacent channel frequency of the TCH frequency of cell cN. Furthermore, the TCH frequency of cell c1 must not be a co-channel or adjacent channel frequency of the BCCH frequency of cell cN. Last but not least, the TCH frequency of cell c1 must not be a co-channel frequency with the TCH frequency of cell cN. Nevertheless, in the last case, the TCH frequency of cell c1 can be an adjacent channel frequency with respect to another TCH frequency of cell cN. In other words, the TCH frequencies of the neighboring cells must not be co-channel frequencies, but can be adjacent channel frequencies.

[0076] The cell proximity conditions serve the purpose of ensuring reliable handover of certain mobile communication participants moving between the area coverage of neighboring cells. When a participant moves away from the antenna (center) of one cell, for example cell c1 in FIG. 3a, during ongoing use of the mobile communication network, its transmission level decreases. As soon as the participant reaches the area of ​​the next neighboring cell, here for example cell cN, the participant must be handed over from cell c1 to cell cN. For this purpose, continuous signal measurements varying between the frequencies of cells c1 and cN are required. If the interference between cells c1 and cN is high, such measurements will fail and, as a consequence, the handover may also fail, which leads to the abortion of the communication. The frequency relationship defined by the cell proximity conditions explained above ensures that the frequency interference between neighboring cells, illustratively c1 and cN in FIG. 3, is kept within acceptable limits so that handover between neighboring cells is reliably guaranteed or the handover failure risk is kept below acceptable limits.

[0077] FIG. 3b exemplarily illustrates further conditions called cell-sector conditions. Here, exemplarily, cells c1, c2, and c3 are located in a certain sector 5. Sector 5 defines, for example, a predefined dense installation area or location, where in this example multiple antennas of different cells c1-c3 are installed close to each other, for example on rooftops, in towns, or in dense urban areas. The cell-sector conditions may define more restrictive frequency relationships than the cell-neighborhood conditions, as explained with respect to FIG. 3a above. This means that the cell-sector conditions according to FIG. 3b may define stricter rules for forbidden frequency relationships between the respective cells (cell pairs) than the cell-neighborhood conditions.

[0078] As exemplarily shown with respect to Fig. 3b, the BCCH and TCH frequencies of one respective cell, here for example cell c1, must not be co-channel (co) or adjacent channel (adj) frequencies of other cells, here for example cells c2 or c3. This means that all frequencies of cell c1 must have a frequency distance of 2 or more to all other frequencies of cells c2 and c3. Thus, the cell-sector condition serves the purpose of avoiding inter-cell interference, which may be significant due to the dense placement and overlap of different cells within sector 5.

[0079] Fig. 4 exemplarily illustrates a further condition to be fulfilled. This further condition is the so-called frequency interference probability condition, which means that a forbidden frequency relationship between the cells of a cell pair in a mobile communication network is defined, for which the frequency interference probability p exceeds a predefined threshold t. In the exemplary diagram of Fig. 4, the cell pair c1 and cN is considered. Also, the co-channel frequency interference probability p c1cN co and adjacent channel frequency interference probability p c1cN adj Furthermore, a threshold value t BCCH and the threshold value t TCH One or both of the co-channel frequency interference probability and the adjacent channel frequency interference probability are equal to or greater than one of the thresholds, and p c1cN co ,p c1cN adj ≧t BCCH ,t TCH , certain frequency relationships are forbidden if one or both of the respective frequency interference probabilities exceed the respective threshold t BCCHThis means that the BCCH frequency of cell c1 must not be the BCCH frequency of cell cN, either for co-channel or adjacent channel frequencies, depending on whether the respective frequency interference probabilities are greater than or equal to the respective threshold t BCCH Similarly, the TCH frequency of cell c1 must not be the TCH frequency of cell cN with respect to either the co-channel frequency or the adjacent channel frequency, depending on whether the respective frequency interference probabilities are greater than or equal to the respective threshold t TCH Finally, the TCH frequency of cell c1 must not be the BCCH frequency of cell cN with respect to either the co-channel frequency or the adjacent channel frequency, depending on whether one or both of the respective frequency interference probabilities are greater than or equal to the respective threshold t TCH Depending on whether the TCH frequency of cell cN is greater than or equal to the co-channel frequency or adjacent channel frequency, similar constraints apply to the exchanged roles of c1 and cN and the interference probability p cNc1 co ,p cNc1 adj applies to.

[0080] In this way, the inter-cell frequency interference must be within acceptable limits between each cell pair, here exemplarily between cell pairs c1 and cN, to prevent adverse effects that could lead to communication failure in the worst case. The overall allocation of frequencies is determined by the probability of interference p co or p adj Even if the frequency interference probability increases, the frequency interference probability between some cells or within a certain region or area of ​​the network after all the unplanned cells are planned and assigned their respective frequencies may exceed a certain threshold t BCCH and t TCH or more is avoided, so that the interference probability can still be kept within acceptable limits.

[0081] 5-9 show exemplary mathematical formulations of the conditions / constraints discussed above. In this regard, the following nomenclature is defined: F={0,1,…,M}: the set of all frequencies C={0,1,…,N}: the set of all cells SEC ⊂ C × C: tuple of cells with sector relation NB ⊂ C × C: a tuple of cells with neighborhood relations. ∀c∈C and ∀f∈F\F(c): x as fixed bit cf BCCH =x cf TCH =0.

[0082] The mathematical formulations in FIGS. 5 to 9 are expressed as so-called Hamiltonian functions, abbreviated as Hamiltonians, and represent the QUBO formulation of qubits as described above.

[0083] The mathematical formulation of Figure 5 formulates the frequency demand condition as described above with respect to Figure 2. The mathematical formulation of Figure 5 represents a binary frequency variable x for each BCCH and TCH frequency that can take on the value "0" or the value "1" (or both with some probability) and is represented as a bit (or qubit as used below) in a quantum-inspired processor. cf BCCH and x cf TCH Thus, the binary frequency variables according to Fig. 5 are constructed as described above with reference to Fig. 2. For each cell c of the set of all cells C in the network 1, and for each frequency f of the frequencies F(c) potentially assigned to each cell c, a respective qubit can be set. If the respective frequency is a BCCH frequency, then the qubit x cf BCCH Otherwise, if each frequency is the TCH frequency, then each qubit x cf TCH are set respectively. Each qubit x cf BCCH , xcf TCH is set to the value "1" if the respective frequency is set accordingly. Otherwise, cf BCCH , x cf TCH is set to the value "0".

[0084] Considering the formulation of the Hamiltonian in Figure 5, the Hamiltonian in the optimal solution must be equal to zero. This is satisfied if and only if the two terms E1 and E2 each become zero. This means that for every cell c in the set C, exactly one frequency f out of the potential frequencies F(c) is assigned as the BCCH frequency and the number of requested frequencies w c is allocated as a TCH frequency. The mathematical formulation according to Fig. 5 allows all unplanned cells, advantageously all cells in network 1, to meet the frequency demand condition described above with reference to Fig. 2.

[0085] The mathematical formulation of Figure 6 formulates the frequency-distance combination condition, as described above with reference to Figure 2. This Hamiltonian is cf BCCH , x cf TCH where we distinguish between different combinations of BCCH or TCH frequencies for co-channel (co) or adjacent channel (adj) frequencies, respectively. Co-channel frequencies are considered by combinations of qubits for the same frequency f, and adjacent channel frequencies are considered by respective combinations of qubits for frequencies f and f+1. Thus, the mathematical formulation of FIG. 6 considers the condition that for all unplanned cells c of the set C and all frequencies f of the potential frequencies F(c), each frequency f in the respective cell c must have a distance ≧2 to other frequencies f, as described above with reference to FIG. 2.

[0086] Considering the Hamiltonian formulation in Figure 6, the Hamiltonian at the optimal solution must be zero. This is only satisfied if each frequency relation (product term combination) obtains the value "0". This means that at least one of each qubit in a product term must obtain the value "0". If instead both qubits in each product term obtain the value "1", then the condition in Figure 6 is not satisfied and the frequency combination distance condition cannot be satisfied.

[0087] The mathematical formulation of the Hamiltonian according to Fig. 7 represents the cell neighborhood condition described above with reference to Fig. 3a. In Fig. 7, all cell pairs c1, c2 are considered that have a neighborhood relation NB in ​​which at least one of the cells c1, c2 is unplanned. For example, here we consider the neighboring cells c1, c2 as described in Fig. 3a. N Also included.

[0088] The Hamiltonian in FIG. 7 is formulated as summed terms for all co-channel (co) frequencies f between cell pair c1, c2 and all adjacent channel (adj) frequencies f, f+1 or f, f−1 between cell pair c1, c2, considering each combination of qubits with respect to the BCCH and / or TCH frequency combination. Thus, the Hamiltonian according to FIG. 7 represents a mathematical formulation of the forbidden frequency relationship as explained with respect to FIG. 3a. Considering the formulation of the Hamiltonian in FIG. 7, the Hamiltonian in the optimal solution must be equal to zero. Here, the same applies as explained with respect to FIG. 6. The Hamiltonian in FIG. 7 is equal to zero only if all product terms of the respective qubit combinations achieve the value “0”. If the frequencies f are assigned such that at least one product term of the respective combination of qubits obtains the value “1”, the condition in FIG. 7 is not met.

[0089] Similar comments as for FIG. 7 apply to the mathematical formulations of FIGS.

[0090] FIG. 8 represents the cell-sector state described above with reference to FIG. 3b. In FIG. 8, all cell pairs c1, c2 with sector relationship SEC are considered. Cells c1, c2 in FIG. 8 represent all cell pairs with sector relationship in the network where at least one of cells c1, c2 is unplanned. For example, cell pairs c1, c2 and c1, c3 in sector 5 described for FIG. 3b are also included here. For all respective cell pairs c1, c2, the Hamiltonian in FIG. 8 is equal to zero if and only if all product terms of the respective qubit combinations achieve the value "0". If the frequency f is assigned such that at least one product term of the respective qubit combinations achieves the value "1", the condition in FIG. 8 is not satisfied.

[0091] FIG. 9 illustrates the frequency interference probability conditions described above with reference to FIG. 4. In FIG. 9, each condition p c1c2 co ,p c1c2 adj ≧t BCCH ,t TCH 9 are considered. The cells c1, c2 in Fig. 9 represent all cell pairs with such a probabilistic relationship in a network where at least one of the cells c1, c2 is unplanned. For example, here we consider the neighboring cells c1, c2 as described for Fig. 4. N The Hamiltonian of Figure 9 is equal to zero if and only if all product terms of each combination of qubits achieve the value "0". If the frequency f is assigned such that at least one product term of each combination of qubits achieves the value "1", then the condition of Figure 9 is not satisfied.

[0092] The mathematical formulation of the Hamiltonian in Fig. 10 represents the calculated frequency interference probability p c1c2 co ,p c1c2 adjThe cells c1, c2 in Fig. 10 represent all cell pairs c1, c2 that have such a probability relationship in a network where at least one of the cells c1, c2 is unplanned. Thus, in the core optimization problem in Fig. 10, for each BCCH and TCH frequency, the condition p c1c2 co ,p c1c2 adj <t BCCH ,t TCH Only the frequency interference probability with is considered. The core optimization problem consists in minimizing the Hamiltonian according to Figure 10 to find the optimized frequency allocation for every cell pair c1, c2 such that the frequency interference between the cells is minimized.

[0093] The Hamiltonian from FIG. 10 contains summed terms, each of which represents a respective frequency interference probability p c1c2 co ,p c1c2 adj by frequency relations involving different combinations of different qubits set as described above. In other words, the Hamiltonian of FIG. 10 is formulated as a stress function that penalizes frequency relations between each cell pair that lead to frequency interference between the respective cells with the calculated frequency interference probability. If each combination of qubits in each frequency relation of the term according to FIG. 10 has a value of “1”, then this frequency combination has a respective frequency interference probability p c1c2 co ,p c1c2 adj Therefore, the mathematical formulation in Fig. 10 can be used to calculate the frequency interference probability p c1c2 co ,p c1c2 adj Add up all the contributions of

[0094] The Hamiltonian of FIG. 10 is typically optimized for all (partially) unplanned cell pairs in the network by a quantum-inspired processor that goes through various settings of values ​​for each qubit and thereby calculates the respective results of the Hamiltonian. The goal in doing so is to find the minimum of the Hamiltonian for each setting of values ​​of the qubits. As soon as the respective minimum of the Hamiltonian of FIG. 10 is found, the respective values ​​of the qubits that lead to this minimum are stored, ultimately defining the optimal frequency allocation with respect to cell interference. Other frequencies that are not candidates for co-channel or adjacent channel frequency interference between the cell pair c1, c2 of FIG. 10 (and therefore have a frequency distance ≧2 with respect to all other cells) can be assigned to the respective cells without further consideration in the Hamiltonian of FIG. 10. Those frequencies are assumed to have a sufficiently high frequency (channel) distance so as not to play a significant role and contribution to the interference between the cells.

[0095] FIG. 11 finally shows the global QUBO formulation of the overall optimization problem, where the partial optimization problems according to FIGS. 5-10 are multiplied by their respective weighting factors A and B and summed to a global optimization problem, which is finally processed by applying a computer-implemented algorithm in a quantum-inspired processor. In this regard, a minimization of the Hamiltonian according to FIG. 10 is performed, which takes into account further optimization constraints formulated in the Hamiltonian according to FIGS. 5-9.

[0096] According to FIG. 11, the core optimization problem of FIG. 10 is multiplied by a weighting factor A, and the sum of other optimization constraint conditions according to FIGS. 5 to 9 is multiplied by a weighting factor B. Using these weighting factors A and B, different weights and focuses can be set for different sub-optimization problems. For example, when the factor B is larger than the factor A, the focus is placed on satisfying the optimization constraint conditions as described with respect to FIGS. 5 to 9. The closer the factor A increases to the factor B, the greater the weight is applied to the optimization of interference by the optimization problem formulated in FIG. 10 described above.

[0097] In certain implementations, it is acceptable for certain constraint conditions to be violated. For example, the constraint condition in FIG. 5 (frequency demand condition) may be violated for certain cells. However, even in these scenarios, if there is at least one cell for which no BCCH frequency is assigned, the result of the algorithm calculation of the method described above may be invalid. In such a case, the constraint condition in FIG. 5 can be split into different constraint conditions that separately follow items E1 and E2 in FIG. 5. In this case, item E1 must be satisfied for all cells to be planned, while item E2 may be violated for certain cells. For example, item E1 can be weighted by factor B, and item E2 can be weighted by factor C, where A < C < B is defined. In this way, item E2 defines a soft constraint condition, and item E1 defines a hard constraint condition. In this way, certain constraint conditions can be flexibly modeled for the actual scenarios and use cases of network planning.

[0098] FIG. 12 shows an exemplary schematic diagram of an algorithm implementing the approach described above. FIG. 12 shows the processing of the method steps and procedures described above considering frequency allocation to unplanned cells 2, which are processed in a decomposed form. This has the advantage, or even is necessary, of processing the described method despite the limited hardware performance of the quantum-inspired processor 7 for solving the optimization problem as described above with respect to FIGS. 5-11. This optimization problem (see in particular FIG. 11) is very complex and must be decomposed into several partial solutions that can be processed iteratively in order to find an optimal solution. At each iteration step, one partial solution is found by the quantum-inspired processor 7.

[0099] The process starts with certain planned cells 6 and unplanned cells 2. If there are no planned cells 6, the process starts with one randomly selected unplanned cell 2. However, if there are one or more already planned cells 6, the process starts with a so-called planned cell criterion in order to decompose the optimization problem. This means that in a first step, the candidates for unplanned cells 2 are sorted / prioritized according to certain constraints or cell relations with respect to the already planned cells 6. For example, first the unplanned cells 2 that have the most cell relations to the already planned cells 6 are considered. The frequency allocation procedure is most affected by the constraints of the already planned cells 6 in terms of potential frequency interference. These unplanned cells 2 are then included first in the subset 8 of unplanned cells 2.

[0100] In a second step, the process continues on a so-called outside-in basis. For this purpose, further candidates for unplanned cells 2 not yet included in the subset 8 of unplanned cells 2 are selected according to certain constraints or cell relations with respect to the cells already included in the subset 8 of unplanned cells 2. These unplanned cells 2 are then also included in the subset 8. This second step is repeated until an upper limit of frequency variables (qubits) that can be calculated at once by the quantum-inspired processor 7 is reached. The constructed subset 8 is then input to an algorithmic procedure in the quantum-inspired processor 7. For example, the quantum-inspired processor 7 according to FIG. 12 is configured to solve an optimization problem by quantum annealing emulation. The quantum-inspired processor 7 applies the mathematical formulation of the global optimization problem according to FIG. 11. The quantum-inspired processor 7 then calculates an optimized solution of the global optimization problem according to FIG. 11 for the subset 8.

[0101] After the algorithm procedure is completed, the final calculated minimum of the global optimization problem according to Fig. 11 is then output from the quantum inspired processor 7 for each subset 8. The cells in the subset 8 are updated to planned cells 6 with the determined optimized frequency allocation. If there are any unplanned cells 2 remaining, the procedure is carried out iteratively for the remaining set of unplanned cells 2 until all cells have been planned and processed. In this case, the global solution for all cells in the communication network is stored. The algorithm then terminates.

[0102] Thus, by applying the computer-implemented algorithmic procedure according to Fig. 12, based on the above implementations and explanations regarding Figs. 1-11, an optimized frequency allocation can be provided for all cells c in the communication network 1. This process is a hybrid method between a pre-processing step (forming the respective subsets 8) and a solution step by using a quantum-inspired processor 7. The procedure according to Fig. 12 considers highly conflicted unplanned cells 2 early in the process to allow maximum freedom in the optimization. This further improves the quality of the solution compared to state-of-the-art approaches.

[0103] The formulation of the optimization problem as a QUBO representation has elegant properties with respect to quantum inspired computing applied here in the processor 7. Today, quantum inspired computing is still reaching significant limitations. However, as computer science increasingly develops towards quantum computing, the approach described herein can be further improved and developed in the future. For example, as quantum computing becomes more and more applicable due to the increase in the size of the underlying optimization problem, the decomposition strategy as described with reference to FIG. 12 can be reduced more and more. This means that the optimization problem can be increasingly treated and computed as a whole without the problem decomposition steps and iterative iterations. Furthermore, as quantum computing becomes more and more applicable, an increasingly increasing number of qubits, an increasingly complex optimization problem, and / or an increasing number of nonlinear constraints can be taken into account by the approach described herein.

[0104] The approach described here is primarily applicable to the allocation of a required number of frequencies to cells in a mobile communication network, but the solution can also be used for allocation problems such as the planning of circuit-separated mobile networks, the planning of code-separated mobile networks, or optimal allocation during operating hours, for example to react to workload spikes.

[0105] The embodiments shown and described herein are merely examples. [Explanation of symbols]

[0106] 1. Communication Network 2 Unplanned cell 3 participants 4 Nearest Neighborhood 5 Sectors 6 Planned cell 7 Quantum-inspired processor 8 Subset of Unplanned Cells a1~aN Antenna C The set of all cells c, c1–cN single cells co-channel adj adjacent channel BCCH Broadcast Channel Frequency TCH Traffic Channel Frequency NB Neighborhood Relationship SEC Sector E1, E2 Frequency demand conditions f frequency p Frequency interference probability t Threshold w number of frequencies requested x Frequency variable

Claims

1. 1. A computer-implemented method for optimizing the allocation of frequencies (f) to cells (C) of a mobile communication network (1), the cells (C) being distributed for communication within the mobile communication network (1), the method comprising: specifying a set of unplanned cells (2) in the mobile communication network (1); specifying, for each unplanned cell (2), a set of frequencies (f) potentially assigned to said unplanned cell (2); calculating a frequency interference probability (p) of selected cell pairs, each cell pair defining a relationship between an unplanned cell (2) and another cell (c) in said mobile communication network (1); formulating stress function terms, each term relating a calculated frequency interference probability (p) of a respective cell pair to a frequency relationship between cells (2,c) of the respective cell pair; determining an optimized allocation of frequencies (f) by selecting for each unplanned cell (2) a subset of frequencies (f) from said respective set of frequencies such that said stress function is minimized by using a quantum inspired processor (7). method.

2. Frequency variable (x cf ), where each frequency variable (x cf ) is associated with one cell (c) in said mobile communication network (1) and with one frequency (f) of said cell (c); The selected frequency variable (x cf formulating the frequency relationship in terms of the stress function as a combination of Using the quantum inspired processor (7), the terms of the stress function are calculated to vary the frequency variables (x cf ) and Further comprising: The method of claim 1.

3. 3. The method according to claim 1 or 2, wherein in terms of the stress function only calculated frequency interference probabilities (p) below a predefined threshold (t) are taken into account.

4. The method comprises the steps of: (1) determining whether each subset of frequencies (f) assigned to an unplanned cell (2) is a designated number (w c 4. The method according to claim 1 , wherein the method is performed taking into account a frequency demand condition that the mobile station has frequencies (f) of 100 MHz to 500 MHz, one frequency (f) being defined as a control channel frequency (BCCH) and another frequency (f) being defined as a traffic channel frequency (TCH).

5. 5. The method according to claim 1, wherein the method is performed taking into account a frequency combination distance condition in which for each unplanned cell (2), a forbidden frequency relationship is defined such that the frequencies (f) of the subset of frequencies (f) assigned to this cell (2) have a frequency relationship with a certain channel distance.

6. 6. The method according to claim 1, wherein the method is carried out taking into account a cell neighbourhood condition in that for cells (c) having a determined neighbourhood relation (4), a forbidden frequency relation between these cells (c) is defined.

7. 7. The method according to claim 1, wherein the method is performed taking into account a frequency interference probability condition in which a forbidden frequency relationship is defined between the cells (2, c) of a cell pair for which the frequency interference probability (p) is above a predefined threshold (t).

8. The frequency relationship distinguishes between a co-channel frequency relationship and an adjacent channel frequency relationship; The co-channel frequency relationship is formulated for frequencies (f) of the same channel (co), and the adjacent channel frequency relationship is formulated for frequencies (f) of adjacent channels (adj).

8. The method according to any one of claims 1 to 7.

9. 9. The method according to claim 1, wherein the frequency relationship distinguishes between a control channel frequency (BCCH) and a traffic channel frequency (TCH).

10. Frequency variable (x cf ), where each frequency variable (x cf ) is associated with one cell (c) in said mobile communication network (1) and with one frequency (f) for said cell (c); The frequency relationship for each condition is plotted against the selected frequency variable (x cf ) in terms of Using the quantum inspired processor (7), the frequency relationship of each condition is calculated and the frequency variable (x cf and setting the 10. The method according to any one of claims 5 to 9.

11. The calculated frequency interference probability (p) of each cell pair is calculated by multiplying the co-channel frequency interference probability (p co ) and adjacent channel frequency interference probability (p adj ) and The co-channel frequency interference probability (p co ) is calculated for the frequency (f) of the same channel (co) between the cells (2,c) of each cell pair, and the adjacent channel frequency interference probability (p adj ) is calculated for the frequency (f) of the adjacent channel (adj) between cells (2,c) of each cell pair, 11. The method according to any one of claims 1 to 10.

12. designating a subset (8) of unplanned cells (2) from said set of unplanned cells (2); performing said method for said subset (8) of unplanned cells (2); updating the cells in the subset to planned cells (6) with the determined optimized frequency (f) allocation; and performing the method iteratively for remaining unplanned cells (2) of the set of unplanned cells (2) until all unplanned cells (2) of the set of unplanned cells (2) have been processed.

12. The method according to any one of claims 1 to 11.

13. A quantum-inspired processor (7), in particular a digital annealing processing unit or a quantum annealing processing unit, configured to perform one or more steps of the method according to any one of the claims 1 to 12.

14. 13. A computer program comprising instructions which, when executed by one or more processors, cause each of said one or more processors to perform one or more steps of the method according to any one of claims 1 to 12.

15. 13. An interface device comprising one or more interfaces to cells (C) of a mobile communication network (1), the cells (C) being distributed for communication within the mobile communication network (1), the interface device being configured to automatically deploy an optimized allocation of frequencies (f) determined by a method according to any one of claims 1 to 12 to the cells (C) of the mobile communication network.

Images