Mobile communication PCI (Peripheral Component Interconnect) decision-making method and device based on improved ant colony algorithm

Abstract

The embodiment of the invention provides a mobile communication PCI (Peripheral Component Interconnect) decision-making method and device based on an improved ant colony algorithm, and the method comprises the steps: obtaining MR (Measurement Report) data of a base station and a real-time network state; the base station MR measurement report data comprises a conflict matrix, a confusion matrix and a modulo 3 interference matrix; extracting a network topology feature vector according to the MR measurement report data of the base station; according to the network topology feature vector and the real-time network state, obtaining a heuristic weight and a pheromone volatilization rate parameter; based on an ant colony algorithm, obtaining an improved ant colony algorithm according to the heuristic weight and the pheromone occurrence rate parameter; based on the quantum revolving door, determining a three-level objective function according to the conflict matrix, the confusion matrix and the modulo 3 interference matrix; and based on an improved ant colony algorithm, generating an optimal PCI allocation scheme according to the conflict matrix, the confusion matrix, the modulo 3 interference matrix, the three-level objective function and a preset constraint condition. The problem of optimization target imbalance is solved, and the reliability of the PCI decision method can be ensured.

Description

TECHNICAL FIELD

[0001] The present document relates to the technical field of mobile communication network optimization, and particularly relates to a mobile communication PCI decision method and device based on an improved ant colony algorithm. BACKGROUND

[0002] In a mobile communication network, the generation mechanism of PCI (Physical Cell Identifier) directly affects network performance and user experience. Its core role is to reduce signal interference by reasonably allocating PCI values, and to ensure normal synchronization and channel decoding of devices.

[0003] The prior art mainly allocates PCI by using a static algorithm (such as a genetic algorithm or an ant colony algorithm) that relies on fixed parameters and artificial rules, or by converting PCI planning into a multi-objective combinatorial optimization problem through a heuristic search algorithm to find an optimal solution.

[0004] However, the static rules and periodic updates in the prior art cannot respond to second-level network changes. If a weighted sum in a heuristic search algorithm is used to forcibly flatten the priority differences of conflict / ambiguity / mod-3 interference, it will lead to unbalanced optimization objectives, and there is a scale bottleneck. SUMMARY

[0005] In view of the above solutions, the present application aims to provide a mobile communication PCI decision method and device based on an improved ant colony algorithm to solve at least one of the above technical problems.

[0006] In a first aspect, one or more embodiments of the present specification provide a mobile communication PCI decision method based on an improved ant colony algorithm, comprising:

[0007] obtaining base station MR measurement report data and real-time network state; the base station MR measurement report data includes a conflict matrix, an ambiguity matrix, and a mod-3 interference matrix;

[0008] based on a graph attention network, extracting a network topology feature vector according to the base station MR measurement report data;

[0009] based on deep reinforcement learning, obtaining heuristic weights and pheromone evaporation rate parameters according to the network topology feature vector and the real-time network state;

[0010] based on an ant colony algorithm, obtaining an improved ant colony algorithm according to the heuristic weights and the pheromone evaporation rate parameters;

[0011] based on a quantum rotation gate, determining a three-level objective function according to the conflict matrix, the ambiguity matrix, and the mod-3 interference matrix;

[0012] Based on the improved ant colony algorithm, according to the conflict matrix, the confusion matrix, the modulo 3 interference matrix, the three-level objective function and the preset constraint condition, an optimal PCI allocation scheme is generated.

[0013] Further, based on a graph attention network, a dynamic topology adjacency relationship model is constructed according to the base station MR measurement report data.

[0014] According to the dynamic topology adjacency relationship model, a dynamic network topology graph is generated.

[0015] According to the dynamic network topology graph, a network topology feature vector is extracted.

[0016] Further, the dynamic topology adjacency relationship model calculation formula is as follows:

[0017]

[0018] Wherein, h ’ i Indicates the updated features of node i, that is, the dynamic network topology graph formed by the adjacency relationship between cells;

[0019] σ represents a nonlinear activation function;

[0020] N represents N cells;

[0021] i represents the number of MRs of cell i;

[0022] j represents the number of MRs of cell j;

[0023] ∝ ij Indicates the attention weight of adjacent node j and cell node i;

[0024] ∝ ij = Softmax (LeakyRELU (W[h i ][h j ]))

[0025] Softmax represents the normalized attention coefficient;

[0026] LeakyRELU is an activation function;

[0027] W is a learnable parameter;

[0028] h i Indicates the node feature;

[0029] h j Indicates the feature of adjacent node j.

[0030] Further, based on the quantum rotation gate, a hierarchical decision model is constructed according to the conflict matrix, the confusion matrix and the modulo 3 interference matrix.

[0031] construct a first layer quantum gate when the number of MRs in the conflict matrix is the lowest;

[0032] construct a second layer quantum gate when the number of MRs in the conflict matrix is the lowest and the number of MRs in the confusion matrix is the lowest;

[0033] construct a third layer quantum gate when the number of MRs in the conflict matrix is the lowest, the number of MRs in the confusion matrix is the lowest, and the number of MRs in the modulo 3 interference matrix is the lowest;

[0034] obtain a three-level target function according to the first layer quantum gate, the second layer quantum gate, and the third layer quantum gate.

[0035] Further, the calculation formula of the three-level target function is as follows:

[0036] The first layer quantum gate is:

[0037]

[0038] wherein, U conflict represents a quantum rotation operation of conflict suppression;

[0039] represents a conflict intensity adjustment factor;

[0040]

[0041] A ij represents the number of conflict MRs of the cell i adjacent area node j;

[0042] B ij represents the number of confusion MRs of the cell i adjacent area node j;

[0043] C ij represents the modulo 3 interference intensity of the cell i adjacent area node j;

[0044] The second layer quantum gate is:

[0045]

[0046] wherein, U confusion represents a quantum controlled non-inversion gate of confusion suppression;

[0047] q A represents a quantum bit register under conflict;

[0048] q B represents a quantum bit register under confusion suppression;

[0049] I C represents a unit matrix, and represents a modulo 3 interference register q CNo operation (keep original state);

[0050] The third layer quantum gate is:

[0051]

[0052] Wherein, U mod3 represents a quantum controlled rotation gate with modulo 3 disturbance compensation;

[0053] q C represents a quantum bit register under modulo 3 disturbance;

[0054] I A represents an identity matrix, indicating that the conflict register state is unchanged;

[0055] I B represents an identity matrix, indicating that the confusion register state is unchanged;

[0056] R z represents a quantum gate rotating around the Z axis, which realizes the accurate control of the quantum state by changing the phase (not the probability amplitude) of the quantum bit;

[0057] represents the rotation angle of the modulo 3 disturbance compensation.

[0058] Further, based on the improved ant colony algorithm, according to the conflict matrix, the confusion matrix and the modulo 3 disturbance matrix, an ant colony is randomly generated;

[0059] Based on the preset constraint condition, the ant colony is processed according to the three-level target function to generate an optimal PCI allocation scheme.

[0060] Further, the optimal PCI allocation scheme calculation formula is as follows:

[0061] PCI i = argmax j |〈j|q i 〉| 2

[0062] Wherein, PCI i represents an optimal PCI allocation scheme;

[0063] j represents the number of MRs of cell j;

[0064] q i represents the superposition of the quantum state of cell i, representing the quantum probability distribution of all possible PCI allocation values (0-1007 integers) of cell i.

[0065] In a second aspect, the embodiments of the present application provide a mobile communication PCI decision device based on an improved ant colony algorithm, comprising:

[0066] an acquisition module configured to acquire base station MR measurement report data and real-time network state, wherein the base station MR measurement report data comprises a conflict matrix, a confusion matrix and a modulo-3 interference matrix;

[0067] a vector extraction module configured to extract a network topology feature vector based on a graph attention network and according to the base station MR measurement report data;

[0068] a parameter optimization module configured to obtain heuristic weight and pheromone evaporation rate parameters based on deep reinforcement learning and according to the network topology feature vector and the real-time network state;

[0069] an improved ant colony algorithm module configured to obtain an improved ant colony algorithm based on an ant colony algorithm and according to the heuristic weight and the pheromone evaporation rate parameters;

[0070] a target module configured to determine a three-level target function based on a quantum rotation gate and according to the conflict matrix, the confusion matrix and the modulo-3 interference matrix;

[0071] a scheme generation module configured to generate an optimal PCI allocation scheme based on the improved ant colony algorithm and according to the conflict matrix, the confusion matrix, the modulo-3 interference matrix, the three-level target function and a preset constraint condition.

[0072] Further, the vector extraction module is configured to construct a dynamic topology adjacency relationship model based on a graph attention network and according to the base station MR measurement report data;

[0073] generate a dynamic network topology graph according to the dynamic topology adjacency relationship model;

[0074] extract a network topology feature vector according to the dynamic network topology graph.

[0075] Further, the vector extraction module is further configured to:

[0076] the dynamic topology adjacency relationship model calculation formula is as follows:

[0077]

[0078] wherein, h ’ i represents the updated feature of node i, i.e. a dynamic network topology graph constituted by the adjacency relationship between cells;

[0079] σ represents a nonlinear activation function;

[0080] N represents N cells;

[0081] i represents the number of MRs of cell i;

[0082] j represents the number of MRs of cell j;

[0083] ∝ ij represents the attention weight of the neighboring node j and the cell node i;

[0084] ∝ ij = Softmax (LeakyRELU (W[h i ][h j ]))

[0085] Softmax represents the normalized attention coefficient;

[0086] LeakyRELU is an activation function;

[0087] W is a learnable parameter;

[0088] h i represents the node feature;

[0089] h j represents the feature of the neighboring node j.

[0090] In a third aspect, the embodiments of the present application provide a storage medium for storing computer executable instructions, characterized in that the computer executable instructions, when executed, implement the steps of the mobile communication PCI decision method based on the improved ant colony algorithm in any one of the first aspect.

[0091] Compared with the prior art, the present application can at least achieve the following technical effects:

[0092] The present application can intelligently extract network topology features through a graph attention network, and combine a deep reinforcement learning dynamic parameter adjustment mechanism to effectively overcome the convergence lag problem of traditional ant colony algorithms in large-scale scenarios, realize efficient optimization of complex networks, implement perception of network change states, adaptively adjust optimization strategies, improve work efficiency, and use three-level target function progressive optimization logic to improve network spectrum efficiency and switching stability. BRIEF DESCRIPTION OF DRAWINGS

[0093] In order to more clearly illustrate the technical solutions in the one or more embodiments of the present specification or the prior art, the drawings needed in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments described in the present specification, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0094] Figure 1 A flowchart of a mobile communication PCI decision method based on an improved ant colony algorithm is provided for one or more embodiments of the present specification;

[0095] Figure 2 A structure diagram of a mobile communication PCI decision device based on an improved ant colony algorithm is provided for one or more embodiments of the present specification. DETAILED DESCRIPTION

[0096] In order for those skilled in the art to better understand the technical solutions in one or more embodiments of the present specification, the technical solutions in one or more embodiments of the present specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of the present specification. Obviously, the described embodiments are only a part of the embodiments of the present specification, rather than all the embodiments. Based on one or more embodiments of the present specification, all other embodiments obtained by those skilled in the art without creative labor should belong to the protection scope of the present document.

[0097] The traditional PCI planning method faces three core problems of insufficient dynamic adaptability, rigid multi-objective optimization, and low large-scale network convergence efficiency in the 5G / 6G ultra-dense networking scenario. In particular, the existing static algorithm (such as genetic algorithm, ant colony algorithm) relies on fixed parameters and artificial rules, and is difficult to respond to real-time changes such as base station expansion and user density fluctuations, resulting in a decline in network conflict suppression capability; at the same time, the weighted summation optimization mode cannot distinguish the priority differences of conflict, confusion and mod3 interference, causing high-weighted targets to mask low-weighted key indicators; in addition, when the traditional method processes a network topology of more than 2000 cells, the convergence speed sharply decreases due to the combination explosion problem, which seriously affects the timeliness of the existing network optimization.

[0098] In view of the above technical problems, the present application proposes a mobile communication PCI decision method based on an improved ant colony algorithm, as shown in Figure 1 The specific steps are as follows:

[0099] Step S1, obtaining base station MR measurement report data and real-time network state; the base station MR measurement report data includes a conflict matrix, a confusion matrix, and a mod3 interference matrix.

[0100] In an embodiment of the present application, the MR data is a report reported by the UE device in the communication process in a timely manner. In each MR data, the main information includes the master cell accessed by the UE in communication and the received adjacent cell information and the corresponding signal strength value.

[0101] For example, the PCI planning decision problem based on MR data is as follows: given N cells, traverse all MR data of these cells to generate N*N matrices, respectively:

[0102] Conflict matrix A = [a ij ] N×N , where if cells i and j are co-frequency, a ijThe value of a is the number of MRs of cell i as the master and j as the neighbor, otherwise a ij The value of a is 0.

[0103] The confusion matrix B = [b ij ] N×N , where if cells i and j are co-frequencies, b ij The value of b is the number of MRs of cells i and j as the neighbor of another cell k, otherwise b ij The value of b is 0.

[0104] The interference matrix c = [c ij ] N×N , where if cells i and j are co-frequencies, c ij The value of c is the number of MRs of cell i as the master and j as the overlapping coverage neighbor of i, otherwise c ij The value of c is 0.

[0105] If cells i and j are assigned the same PCI value, the number of conflicts increases a ij +a ji , the number of confusions increases b ij +b ji , and if cells i and j are assigned the same value of PCI mod 3, the number of mod 3 interference increases c ij +c ji .

[0106] In the present application, since the measurement report is reported at a fixed time interval, the distribution of the MR data can basically reflect the distribution of the traffic volume. At the same time, since the MR data is comprehensive in time and space, the number of conflicts, confusions and mod 3 interference of the MR data is used to reflect the PCI index in the network, which is more accurate.

[0107] In step S2, a network topology feature vector is extracted based on a graph attention network according to the base station MR measurement report data.

[0108] In the embodiments of the present application, a dynamic topology adjacency relationship model is constructed based on a graph attention network according to the base station MR measurement report data; a dynamic network topology graph is generated according to the dynamic topology adjacency relationship model; and a network topology feature vector is extracted according to the dynamic network topology graph.

[0109] Specifically, the graph attention network (GAT) is used to dynamically model the neighbor relationship, and the calculation formula is as follows:

[0110] ∝ ij = Softmax (LeakyRELU (W[h i ][h j ]))

[0111] Where, ∝ ijdenotes the attention weight of the neighboring node j and the cell node i;

[0112] h i is the node feature, W is a learnable parameter, LeakyRELU is an activation function, and Softmax is a normalized attention coefficient

[0113] Node feature update:

[0114]

[0115] wherein h ’ i denotes the updated feature of the node i, i.e., a dynamic network topology graph formed by the adjacency relationship between cells;

[0116] σ denotes a nonlinear activation function;

[0117] N denotes N cells;

[0118] i denotes the number of MRs of cell i;

[0119] j denotes the number of MRs of cell j.

[0120] Step S3, based on deep reinforcement learning, obtaining heuristic weight and pheromone evaporation rate parameters according to the network topology feature vector and the real-time network state.

[0121] In the embodiments of the present application, based on the network topology feature vector, in each optimization period, a deep reinforcement learning (SAC) agent is introduced to dynamically adjust the heuristic weight and pheromone evaporation rate parameters of the ant colony algorithm according to the real-time network state (user density fluctuation, base station expansion event);

[0122] Specifically, based on the real-time network state, a deep reinforcement learning (Soft Actor-Critic, SAC) is introduced to dynamically adjust the heuristic weight of the ant colony algorithm, and the heuristic weight (β) and pheromone evaporation rate (ρ) of ACO are dynamically adjusted by optimizing the state. The Soft Actor-Critic (SAC) algorithm is adopted, and the state space in the model is:

[0123] s t = [H t ,τ t ,F history ]

[0124] wherein s t denotes the state space of the deep reinforcement learning (SAC) agent at time t, which is essentially a feature vector that integrates network dynamics and optimization history, and its function is to provide decision basis for SAC and dynamically adjust the parameters of the ant colony algorithm;

[0125] Ht ∈R n×m is the GAT embedding matrix of all cells, representing the topology and interference characteristics; n is the number of cells, and m is the number of countable PCI values;

[0126] τ t ∈R n×m is the pheromone matrix, recording the preference degree of each cell to each PCI;

[0127] F history is the historical interference statistics.

[0128] The action space is:

[0129] Δβ∈[-0.1,0.1],Δρ∈[-0.05,0.05]

[0130] The update rule is:

[0131] β t+1 = clip(β t +Δβ,1,5),ρ t+1 = clip(ρ t +Δρ,0.1,0.5)

[0132] Where clip is a limit function.

[0133] In this application, by real-time sensing of network state changes (such as user density fluctuations, base station expansion), the SAC agent adaptively adjusts the optimization strategy by maximizing the expected reward, greatly reducing the need for manual intervention, providing reliable technical support for 5G / 6G ultra-dense networking and future network intelligent evolution, while solving the problem of low convergence efficiency caused by fixed parameters.

[0134] Step S4, based on the ant colony algorithm, according to the heuristic weight and the pheromone emission rate parameter, an improved ant colony algorithm is obtained.

[0135] In the embodiments of the present application, the ant colony algorithm is optimized by using the heuristic weight and the pheromone emission rate parameter, and an improved ant colony algorithm is obtained.

[0136] Step S5, based on the quantum rotation gate, according to the conflict matrix, the confusion matrix and the modulo 3 interference matrix, a three-level target function is determined.

[0137] In the embodiment of the present application, based on the quantum rotation gate, a hierarchical decision model is established according to the conflict matrix, the confusion matrix and the modulo 3 interference matrix; when the number of MRs in the conflict matrix is the lowest, a first layer quantum gate is constructed; when the number of MRs in the conflict matrix is the lowest and the number of MRs in the confusion matrix is the lowest, a second layer quantum gate is constructed; when the number of MRs in the conflict matrix is the lowest, the number of MRs in the confusion matrix is the lowest and the number of MRs in the modulo 3 interference matrix is the lowest, a third layer quantum gate is constructed; and a three-level target function is obtained according to the first layer quantum gate, the second layer quantum gate and the third layer quantum gate.

[0138] Specifically, for the priority difference of conflict, confusion and modulo 3 interference, a quantum rotation gate sequence is constructed:

[0139] First, quantum state encoding and initialization are performed.

[0140] The conflict matrix A, the confusion matrix B and the modulo 3 interference matrix C after input normalization processing are mapped to quantum amplitudes:

[0141]

[0142] Where ψ ij represents the probability amplitude of converting the three types of interference indicators (conflict A ij , confusion B ij , modulo 3 interference C ij ) of cell i and neighboring area j into quantum state |ψ ij 〉, which is used for subsequent quantum optimization;

[0143] A ij represents the number of conflict MRs of cell i neighboring area node j;

[0144] B ij represents the number of confusion MRs of cell i neighboring area node j;

[0145] C ij represents the modulo 3 interference intensity of cell i neighboring area node j;

[0146] The output topology embedding h i of the graph attention network is generated to generate a quantum pheromone tensor:

[0147]

[0148] Where τ ijk represents generating the topology correlation strength memory between cell i, neighboring area j and PCI candidate value k, as a dynamic guide factor of the ant colony algorithm;

[0149] k represents another cell;

[0150] W is a learnable parameter;

[0151] h i denotes the characteristics of the node;

[0152] h j denotes the characteristics of the neighboring node j.

[0153] Then, dynamic priority quantum gate control is performed.

[0154] The conflict MR number is minimized, and the first layer quantum gate is constructed:

[0155]

[0156] wherein U conflict (θ) denotes a conflict suppression quantum rotation operation;

[0157] θ denotes a conflict intensity adjustment factor;

[0158] The phase difference of high conflict PCI can be amplified by the first layer quantum gate, and its selected probability is suppressed.

[0159] The second layer quantum gate is constructed to ensure that the confusion is minimized under the premise of conflict priority, that is, when the conflict is serious (q A =1), the confusion PCI value is forced to jump; the formula is as follows:

[0160]

[0161] wherein U confusion denotes a confusion suppression quantum controlled NOT gate;

[0162] q A denotes a quantum bit register under conflict;

[0163] q B denotes a quantum bit register under confusion suppression;

[0164] I C denotes a unit matrix, and q C denotes a no-operation (maintaining the original state);

[0165] The third layer quantum gate is constructed by optimizing the modulo three interference:

[0166]

[0167] wherein U mod3 denotes a modulo 3 interference compensation quantum controlled rotation gate;

[0168] q C denotes a quantum bit register under modulo 3 interference;

[0169] I Adenotes the identity matrix, and denotes that the conflict register state is unchanged;

[0170] I B denotes the identity matrix, and denotes that the confusion register state is unchanged;

[0171] R z denotes the quantum gate of rotating around the Z axis;

[0172] denotes the rotation angle of the modulo 3 interference compensation.

[0173] The application innovatively proposes a quantum-inspired hierarchical priority control architecture, strictly follows the progressive optimization logic of "conflict elimination → confusion suppression → modulo 3 interference optimization", breaks through the target coupling limitation of the traditional weighted summation method, and significantly improves the network spectrum efficiency and switching stability.

[0174] In step S6, based on the improved ant colony algorithm, an optimal PCI allocation scheme is generated according to the conflict matrix, the confusion matrix, the modulo 3 interference matrix, the three-level objective function and a preset constraint condition.

[0175] In the embodiment of the application, based on the improved ant colony algorithm, an ant colony population is randomly generated according to the conflict matrix, the confusion matrix and the modulo 3 interference matrix; based on a preset constraint condition, the ant colony population is processed according to the three-level objective function to generate an optimal PCI allocation scheme.

[0176] Specifically, in combination with the ant colony algorithm, an ant colony population is first randomly generated: Q ants ={q1,...,q M}

[0177] Then, a quantum gate operation sequence is executed to generate a quantum ant colony population:

[0178]

[0179] wherein Q ′ ants denotes the quantum ant colony population;

[0180] k denotes the execution order number of the quantum gate operation (k = 1, 2, 3);

[0181] U k denotes the quantum unitary transformation matrix corresponding to the priority k;

[0182] θ k denotes a dynamically calculated quantum rotation angle parameter;

[0183] Q ants denotes a random ant colony population.

[0184] The candidate solution is obtained by quantum measurement, and the optimal PCI allocation scheme is obtained under the constraint conditions, wherein the constraint conditions include but are not limited to the PCI value range constraint (integer value of 0-1007); the uniqueness constraint of the same frequency cell (the PCI of the same frequency adjacent cell cannot be repeated); the interference avoidance constraint of module 3 (the PCI of the adjacent cell needs to be differentiated in module 3); the topology dependent constraint (the PCI of the high interference associated cell is forced to be isolated); and the calculation formula of the optimal PCI allocation scheme is:

[0185] PCI i = argmax j |〈j|q i 〉| 2

[0186] Wherein, PCI i represents the optimal PCI allocation scheme of the current iteration;

[0187] q i represents the superposition of the state of cell i, which represents the quantum probability distribution of all possible PCI allocation values (integer of 0-1007) of cell i.

[0188] The quantum optimization output result PCI i is obtained, and the field solution space is constructed to optimize and verify the optimal PCI allocation scheme:

[0189]

[0190] Wherein, epoch represents the number of algorithm iterations;

[0191] An evaluation function is constructed again:

[0192]

[0193] Wherein, δ represents the conflict cost function;

[0194] ζ represents the confusion cost function;

[0195] ξ represents the module 3 interference cost function.

[0196] Finally, the pheromone updating mechanism is performed:

[0197]

[0198] Wherein, is the GAT feature weighted heuristic information, and ρ∈(0, 1) is dynamically adjusted by DRL.

[0199] ρ represents the pheromone evaporation coefficient.

[0200] For example, first, the territory relationship of each cell (network topology data (cell location, coverage)) is acquired to perform parameter initialization, m ants are randomly placed on n cells, and the number of cycles N is determined to be N+1; then the ant k selects a road according to a specified function, and it is determined whether the ant k has traversed all cells, if not, the ant k selects a road according to the specified function again; if it has been traversed, it is determined whether all ants (k>m) are completed, if not, the next ant is replaced; if all ants are completed, the pheromone is updated; after the pheromone is updated, it is determined whether the termination condition (including reaching the maximum number of cycles N_max or the convergence condition) is met, if yes, the optimal PCI allocation scheme is output; otherwise, the number of cycles is increased and the loop continues.

[0201] The embodiment of the application provides a mobile communication PCI decision device based on an improved ant colony algorithm, as shown in the figure, comprising, Figure 2

[0202] The acquisition module 101 is configured to acquire base station MR measurement report data and real-time network state, wherein the base station MR measurement report data comprises a conflict matrix, a confusion matrix and a modulo 3 interference matrix.

[0203] The vector extraction module 102 is configured to extract a network topology feature vector based on a graph attention network according to the base station MR measurement report data.

[0204] The parameter optimization module 103 is configured to obtain heuristic weight and pheromone evaporation rate parameters based on deep reinforcement learning according to the network topology feature vector and the real-time network state.

[0205] The improved ant colony algorithm module 104 is configured to obtain an improved ant colony algorithm based on an ant colony algorithm according to the heuristic weight and the pheromone evaporation rate parameters.

[0206] The target module 105 is configured to determine a three-level target function based on a quantum rotation gate according to the conflict matrix, the confusion matrix and the modulo 3 interference matrix.

[0207] The scheme generation module 106 is configured to generate an optimal PCI allocation scheme based on the improved ant colony algorithm according to the conflict matrix, the confusion matrix, the modulo 3 interference matrix, the three-level target function and a preset constraint condition.

[0208] Further, the vector extraction module is configured to:

[0209] construct a dynamic topology adjacency relationship model based on a graph attention network according to the base station MR measurement report data;

[0210] generate a dynamic network topology graph according to the dynamic topology adjacency relationship model. ​

[0211] According to the dynamic network topology graph, a network topology feature vector is extracted.

[0212] Further, the vector extraction module is further configured to:

[0213] The dynamic topology adjacency relationship model calculation formula is as follows:

[0214]

[0215] wherein h ’ i represents the updated feature of the node i, that is, the dynamic network topology graph formed by the adjacency relationship between cells;

[0216] σ represents a nonlinear activation function;

[0217] N represents N cells;

[0218] i represents the number of MRs of cell i;

[0219] j represents the number of MRs of cell j;

[0220] ∝ ij represents the attention weight of the adjacent node j and the cell node i;

[0221] ∝ ij = Softmax(LeakyRELU(W[h i ][h j ]))

[0222] Softmax represents a normalized attention coefficient;

[0223] LeakyRELU is an activation function;

[0224] W is a learnable parameter;

[0225] h i represents a node feature;

[0226] h j represents a feature of the adjacent node j.

[0227] The embodiments of the application provide a storage medium for storing computer executable instructions, characterized in that the computer executable instructions, when executed, implement the steps of the mobile communication PCI decision method based on the improved ant colony algorithm in any one of the above embodiments.

[0228] It should be noted that the embodiment of the storage medium in the specification is based on the same inventive concept as the embodiment of the mobile communication PCI decision method based on the improved ant colony algorithm in the specification, and therefore the specific implementation of this embodiment can be referred to the foregoing implementation of the corresponding mobile communication PCI decision method based on the improved ant colony algorithm, and the repeated parts will not be described herein.

[0229] The above describes specific embodiments of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different than the order in which the actions or steps are recited in the embodiments, and still achieve desirable results. In addition, the processes depicted in the figures do not necessarily require the particular order shown, or sequential order, to achieve the desired results. In certain implementations, multitasking and parallel processing can be advantageous or necessary.

[0230] In the 1930s, it was clear to distinguish whether an improvement in a technology was in hardware (e.g., improvement in circuit structure of diodes, transistors, switches, etc.) or in software (e.g., improvement in method flow). However, as technology has evolved, many improvements in method flow today can be considered as direct improvements in hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented by a hardware entity module. For example, a programmable logic device (PLD) (e.g., a field programmable gate array (FPGA)) is an integrated circuit whose logic function is determined by user programming of the device. A digital system is "integrated" on a PLD by the designer programming it, rather than by asking a chip manufacturer to design and fabricate a custom integrated circuit chip. Moreover, instead of manually fabricating an integrated circuit chip, this programming is now mostly implemented by "logic compiler" software, which is similar to software compilers used in program development, and the original code to be compiled is written in a specific programming language, called a hardware description language (HDL), of which there are many, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc., the most commonly used being VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. It should be clear to those skilled in the art that, by simply logically programming a method flow in one of the above hardware description languages and programming it into an integrated circuit, a hardware circuit implementing the logical method flow can be easily obtained.

[0231] The controller can be implemented in any suitable way, for example, the controller can take the form of a microprocessor or processor and a computer readable medium storing computer readable program code, e.g. software or firmware, executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller and an embedded microcontroller, examples of which include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20 and Silicone Labs C8051F320, the memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also know that, in addition to being implemented in pure computer readable program code, the controller can equally well be implemented to perform the same functions using logic gates, switches, an application specific integrated circuit, a programmable logic controller and an embedded microcontroller, etc. by means of a logical programming of the method steps. The controller can thus be considered as a hardware component, and the means comprised therein for performing the various functions can be considered as structures within the hardware component. Alternatively, the means for performing the various functions can even be considered as both a software module implementing the method and a structure within the hardware component.

[0232] The systems, apparatuses, modules or units illustrated by the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0233] For the sake of description, the above apparatuses are described in various units with functions respectively. Of course, the functions of the units can be implemented in one or more software and / or hardware in implementing the embodiments of the present specification.

[0234] Those skilled in the art will understand that one or more embodiments of the present specification can be provided as a method, a system or a computer program product. Therefore, one or more embodiments of the present specification can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present specification can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROMs, optical storage devices, etc.) containing computer usable program code.

[0235] The specification is presented with reference to flow diagrams and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the specification. 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, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing element or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 The flow diagrams and / or block diagrams in the specification can present a method, apparatus or computer program product according to embodiments of the specification. Flow diagrams and / or block diagrams can also present a method, apparatus or computer program product to achieve functions specified in flow diagrams and / or block diagrams block or blocks. Figure 1 The flow diagrams and / or block diagrams in the specification can present a method, apparatus or computer program product according to embodiments of the specification. Flow diagrams and / or block diagrams can also present a method, apparatus or computer program product to achieve functions specified in flow diagrams and / or block diagrams block or blocks.

[0236] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 The flow diagrams and / or block diagrams in the specification can present a method, apparatus or computer program product according to embodiments of the specification. Flow diagrams and / or block diagrams can also present a method, apparatus or computer program product to achieve functions specified in flow diagrams and / or block diagrams block or blocks. Figure 1 The flow diagrams and / or block diagrams in the specification can present a method, apparatus or computer program product according to embodiments of the specification. Flow diagrams and / or block diagrams can also present a method, apparatus or computer program product to achieve functions specified in flow diagrams and / or block diagrams block or blocks.

[0237] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flow diagrams and / or block diagrams block or blocks. Figure 1 The flow diagrams and / or block diagrams in the specification can present a method, apparatus or computer program product according to embodiments of the specification. Flow diagrams and / or block diagrams can also present a method, apparatus or computer program product to achieve functions specified in flow diagrams and / or block diagrams block or blocks. Figure 1 The flow diagrams and / or block diagrams in the specification can present a method, apparatus or computer program product according to embodiments of the specification. Flow diagrams and / or block diagrams can also present a method, apparatus or computer program product to achieve functions specified in flow diagrams and / or block diagrams block or blocks.

[0238] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0239] The memory can include non-persistent memory and / or persistent memory, such as flash memory, read-only memory (ROM), and / or volatile or non-volatile random access memory (RAM), among others. The memory is an example of computer-readable media.

[0240] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0241] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusion, such that processes, methods, articles or devices that include a series of elements not only include those elements, but also include other elements not explicitly listed, or inherent to such processes, methods, articles or devices. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or device that includes the element.

[0242] One or more embodiments of the present specification can be described in the general context of computer-executable instructions being executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform particular tasks or implement particular abstract data types. One or more embodiments of the present specification can also be practiced in a distributed computing environment, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0243] Each embodiment in the present specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other, and each embodiment focuses on the difference from other embodiments. In particular, for system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiment.

[0244] The above merely provides the example of the present document and is not intended to limit the present document. For those skilled in the art, the present document can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present document shall be included in the scope of claims of the present document.

Claims

1. A mobile communication PCI decision method based on an improved ant colony algorithm, characterized in that, include: Acquire base station MR measurement report data and real-time network status; The base station MR measurement report data includes a collision matrix, a confusion matrix, and a modulo-3 interference matrix. Based on graph attention networks, network topology feature vectors are extracted according to the base station MR measurement report data; Based on deep reinforcement learning, heuristic weights and pheromone evaporation rate parameters are obtained according to the network topology feature vector and the real-time network state. Based on the ant colony algorithm, an improved ant colony algorithm is obtained according to the heuristic weights and the pheromone evaporation rate parameters. Based on the quantum rotating gate, a three-level objective function is determined according to the conflict matrix, the confusion matrix, and the modulo 3 interference matrix; Based on the improved ant colony algorithm, an optimal PCI allocation scheme is generated according to the conflict matrix, the confusion matrix, the modulo 3 interference matrix, the three-level objective function, and the preset constraints.

2. The method according to claim 1, characterized in that, The graph attention-based network extracts network topology feature vectors based on the base station MR measurement report data, including: Based on graph attention networks, a dynamic topological adjacency relationship model is constructed according to the base station MR measurement report data; Based on the dynamic topological adjacency relationship model, a dynamic network topology graph is generated; Based on the dynamic network topology graph, extract the network topology feature vector.

3. The method according to claim 2, characterized in that, The calculation formula for the dynamic topological adjacency relationship model is as follows: in, This represents the updated characteristics of node i, i.e., the dynamic network topology graph formed by the adjacency relationships between cells; Represents a nonlinear activation function; N represents N cells; i represents the number of MRs in cell i; j represents the number of MRs in cell j; This represents the attention weights of neighboring node j and cell node i; Softmax represents the normalized attention coefficient; LeakyRELU is the activation function; W is a learnable parameter; h i Represent node characteristics; h j This represents the characteristics of neighboring node j.

4. The method according to claim 1, characterized in that, The determination of the third-level objective function based on the quantum rotation gate, according to the conflict matrix, the confusion matrix, and the modulo-3 interference matrix, includes: Based on the quantum rotating gate, a hierarchical decision-making model is established according to the conflict matrix, the confusion matrix, and the modulo-3 interference matrix. When the MR number in the conflict matrix is ​​at its lowest, the first layer of quantum gates is constructed. When the MR number in the conflict matrix is ​​the lowest and the MR number in the confusion matrix is ​​the smallest, a second quantum gate is constructed; When the MR number in the conflict matrix is ​​the lowest, the MR number in the confusion matrix is ​​the lowest, and the MR number in the modulo-3 interference matrix is ​​the lowest, a third quantum gate is constructed. Based on the first layer of quantum gates, the second layer of quantum gates, and the third layer of quantum gates, a three-level objective function is obtained.

5. The method according to claim 4, characterized in that, The formula for calculating the third-level objective function is as follows: The first layer of quantum gates is: in, Quantum rotation operations representing conflict suppression; This represents the conflict intensity adjustment factor; This represents the number of conflicting MRs in neighboring cell j of cell i; This represents the number of confused MRs in neighboring node j of cell i; This represents the modulo-3 interference intensity of neighboring node j in cell i; The second layer of quantum gates is: in, Represents a quantum controlled NOT gate for confusion suppression; A register representing a qubit in the event of a collision; This represents a qubit register under confusion suppression. Represents the identity matrix, and represents the modulo-3 interference register. No operation (remain in the same state); The third quantum gate is: in, Represents a quantum-controlled rotating gate with modulo-3 interference compensation; This represents a quantum bit register under modulo-3 interference. Represents the identity matrix, indicating that the state of the conflict register remains unchanged; Represents the identity matrix, indicating that the state of the confusion register remains unchanged; A quantum gate representing rotation about the Z-axis; This indicates the rotation angle for modulo 3 interference compensation.

6. The method according to claim 1, characterized in that, The improved ant colony algorithm, based on the conflict matrix, confusion matrix, modulo 3 interference matrix, third-level objective function, and preset constraints, generates the optimal PCI allocation scheme by including: Based on the improved ant colony algorithm, an ant population is randomly generated according to the conflict matrix, the confusion matrix, and the modulo 3 interference matrix; Based on preset constraints, the ant population is processed according to the three-level objective function to generate the optimal PCI allocation scheme.

7. The method according to claim 6, characterized in that, The formula for calculating the optimal PCI allocation scheme is as follows: in, This represents the optimal PCI allocation scheme; j represents the number of MRs in cell j; Let represent the quantum state superposition of cell i, and let represent the quantum probability distribution of all possible PCI assignment values ​​(integers from 0 to 1007) for cell i.

8. A mobile communication PCI decision-making device based on an improved ant colony algorithm, characterized in that... include: The acquisition module is used to acquire base station MR measurement report data and real-time network status; The base station MR measurement report data includes a collision matrix, a confusion matrix, and a modulo-3 interference matrix. The vector extraction module is used to extract network topology feature vectors based on the base station MR measurement report data, using a graph attention network. The parameter optimization module is used to obtain heuristic weights and pheromone evaporation rate parameters based on deep reinforcement learning, according to the network topology feature vector and the real-time network state. An improved ant colony algorithm module is used to obtain an improved ant colony algorithm based on the ant colony algorithm, according to the heuristic weights and the pheromone evaporation rate parameters. The objective module is used to determine a three-level objective function based on the quantum rotating gate, according to the conflict matrix, the confusion matrix, and the modulo-3 interference matrix. The scheme generation module is used to generate an optimal PCI allocation scheme based on an improved ant colony algorithm, according to the conflict matrix, the confusion matrix, the modulo 3 interference matrix, the three-level objective function, and preset constraints.

9. The apparatus according to claim 8, characterized in that... The vector extraction module is configured as follows: Based on graph attention networks, a dynamic topological adjacency relationship model is constructed according to the base station MR measurement report data; Based on the dynamic topological adjacency relationship model, a dynamic network topology graph is generated; Based on the dynamic network topology graph, extract the network topology feature vector.

10. The apparatus according to claim 9, characterized in that... The vector extraction module is also configured to: The calculation formula for the dynamic topological adjacency relationship model is as follows: in, This represents the updated characteristics of node i, i.e., the dynamic network topology graph formed by the adjacency relationships between cells; Represents a nonlinear activation function; N represents N cells; i represents the number of MRs in cell i; j represents the number of MRs in cell j; This represents the attention weights of neighboring node j and cell node i; Softmax represents the normalized attention coefficient; LeakyRELU is the activation function; W is a learnable parameter; h i Represent node characteristics; h j This represents the characteristics of neighboring node j.

11. A storage medium for storing computer-executable instructions, characterized in that, When executed, the computer-executable instructions implement the steps of the mobile communication PCI decision method based on the improved ant colony algorithm as described in any one of claims 1-7.

Images