A heuristic-based cell pci planning method and system

Abstract

The application provides a heuristic-based cell PCI planning method and system, relates to the technical field of PCI distribution and planning in a mobile communication network, and comprises the steps of multi-objective optimization model establishment, data acquisition and initialization, initial distribution, problem decomposition, parallel optimization and the like. Through the construction of an accurate multi-objective optimization model and the design of an efficient decomposition-parallel system solution framework, and in combination with a heuristic algorithm for quickly generating an initial feasible solution, a cell interaction graph model constructed based on each cell and its neighboring relationship data is segmented by using a graph segmentation algorithm, the multi-objective optimization problem is decomposed into multiple sub-optimization problems and solved by using a multi-objective evolutionary algorithm, the conflict and interference problems among neighboring cells can be effectively solved, the global optimization conflict, confusion and modulo 3 interference are solved while ensuring the solution efficiency and coping with a super-large-scale network, and finally a high-quality globally optimized PCI planning scheme is output, so that the overall network performance and user experience are improved.

Description

Technical Field

[0001] This invention relates to the field of PCI allocation and planning technology in mobile communication networks, and particularly to a heuristic cell PCI planning method and system. Background Technology

[0002] In mobile communication systems, the Physical Cell Identifier (PCI) is a key parameter used to distinguish different cells. Due to limited PCI resources, PCI allocation in densely deployed network scenarios is prone to conflicts and interference. Existing PCI planning methods often employ simple greedy algorithms, but these tend to ignore the complex interference relationships between cells, failing to effectively address conflicts and interference between neighboring cells. Especially in large-scale, high-density network environments, such methods struggle to achieve global optimization, leading to high levels of interference, confusion, and conflicts in the network, primarily: conflict MR (Measurement Report) counts, confusion MR counts, and modulo-3 interference MR counts, further impacting key performance indicators such as user rate and handover success rate. Summary of the Invention

[0003] To address the problems of existing PCI planning methods, such as neglecting conflicts and interference between neighboring cells, difficulty in achieving global optimization, high levels of interference, confusion, and conflict, and inability to adapt to large-scale networks, this paper proposes a heuristic-based PCI planning method and system. This method constructs an accurate multi-objective optimization model and designs an efficient decomposition-parallel system solution framework. After rapidly generating initial feasible solutions using heuristic algorithms, a graph segmentation algorithm is used to segment the cell interaction graph model constructed based on data from each cell and its neighboring relationships. This decomposes the multi-objective optimization problem into multiple sub-optimization problems, which are then solved using a multi-objective evolutionary algorithm. This approach not only fully considers the complex interference relationships between cells and effectively solves conflicts and interference between neighboring cells, but also globally optimizes conflicts, confusion, and modulo 3 interference while ensuring solution efficiency and handling ultra-large-scale networks. Ultimately, it outputs a high-quality, implementable, globally optimized PCI planning scheme, thereby improving overall network performance and user experience.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] The present invention provides a heuristic-based cell PCI planning method in its first aspect, comprising the following steps: S0. Based on the requirements of PCI planning in the communication network, with the objective of minimizing the number of collision MRs, confusion MRs, and modulo 3 interference MRs in the communication network, and at least using PCI allocation uniqueness constraints, collision avoidance constraints, confusion avoidance constraints, and modulo 3 interference avoidance constraints as constraints, a multi-objective optimization model is constructed; S1. Read the data of each cell and its neighbor relationships, and initialize the PCI of all cells to an unallocated state; S2. Using the data of each cell and its neighbor relationships as input to the multi-objective optimization model, an initial allocation is performed using a heuristic algorithm to generate an initial PCI allocation scheme that satisfies the constraints; S3. Based on the data of each cell and its neighbor relationships, a cell interaction graph model is constructed, and a graph partitioning algorithm is used to partition the cell interaction graph model so that the multi-objective optimization problem is decomposed into multiple sub-optimization problems; S4. For each sub-optimization problem, the multi-objective optimization model is solved in parallel using a multi-objective evolutionary algorithm in conjunction with the initial PCI allocation scheme to obtain the optimal set of PCI allocation schemes.

[0006] The present invention provides a preferred embodiment in its first aspect. In step S2, the heuristic algorithm employs a greedy algorithm. The initial allocation process includes: calculating the adjacency degree of each cell based on proximity relationships; sorting the cells by their adjacency degrees, starting with the cell with the highest adjacency degree, and allocating PCIs to each cell in descending order of adjacency degree, ensuring that the PCIs of each cell are different from those of all its neighboring cells. Five core constraints are precisely defined, providing a clear optimization objective and search space for the algorithm. The adjacency-based greedy algorithm quickly generates high-quality initial feasible solutions, significantly accelerating the convergence of subsequent optimization processes.

[0007] In its first aspect, this invention provides a preferred embodiment. Step S3 involves constructing a cell interaction graph model based on data about each cell and its neighboring relationships. Specifically, this includes: treating each cell as a node in the graph; establishing an edge between nodes corresponding to any two adjacent cells based on the input neighboring relationship data; and assigning corresponding weights to this edge based on the input collision MR number, confusion MR number, and modulo 3 interference MR number, thereby forming a weighted cell interaction graph model. By constructing this weighted interaction graph model, complex network interference relationships are transformed into a computable and analyzable graph structure, providing a reliable foundation for subsequent segmentation.

[0008] In its first aspect, this invention provides a preferred solution. In step S3, the graph segmentation algorithm specifically employs the Louvain algorithm. By iteratively optimizing the modularity index, the cell interaction graph model is automatically segmented into multiple internally connected subgraphs, thereby decomposing the multi-objective optimization problem into sub-optimization problems corresponding to each subgraph. The application of the Louvain algorithm automatically identifies highly cohesive subgraphs in the network, achieving intelligent decomposition and optimization of complex problems, fundamentally reducing the scale and difficulty of the solution.

[0009] The present invention provides a preferred embodiment in its first aspect. In step S4, the multi-objective evolutionary algorithm employs the NSGA-II algorithm. The solution process includes: initializing the population under the premise of satisfying constraints; evaluating the fitness of individuals and performing non-dominated sorting according to the objective function in the multi-objective optimization model; calculating the crowding distance of individuals to maintain population diversity; generating offspring populations through selection, crossover, and mutation operations; merging the parent and offspring populations and selecting a new generation population; iterative evolution until termination; wherein, when initializing the population, the PCI allocation scheme of the corresponding subgraph in the initial PCI allocation scheme is included as a complete individual in the initial population. The NSGA-II algorithm is used in parallel to find the optimal solution in each subgraph, guided by the initial solution, to efficiently obtain the optimal solution set that is balanced among multiple competing objectives.

[0010] In a first aspect, this invention provides a preferred embodiment in which, in the step of evaluating individual fitness and performing non-dominated ranking based on the objective function in the multi-objective optimization model, fitness is specifically evaluated by determining the Pareto rank of each individual; in step S4, the obtained set of optimal PCI allocation schemes is: the set of Pareto-optimal PCI allocation schemes in each subgraph. By explicitly using the Pareto rank as the fitness criterion and outputting the Pareto-optimal solution set, it ensures that the result is theoretically the best trade-off between multiple objectives that cannot be further improved.

[0011] In a first aspect, this invention provides a preferred embodiment in which, in the step of generating a progeny population through selection, crossover, and mutation operations, the selection operation is specifically based on a tournament selection mechanism, selecting individuals with fitness and crowding levels both above a preset threshold for subsequent crossover and mutation operations. By integrating Pareto rank and crowding comparison in evolutionary selection, diversity is effectively maintained while promoting population evolution, avoiding premature convergence.

[0012] The present invention provides a preferred embodiment in its first aspect. The heuristic cell PCI planning method further includes: S5. fusing the set of optimal PCI allocation schemes to generate candidate global schemes, and performing conflict and interference verification and adjustment on cells with proximity relationships across subgraph boundaries generated by subgraph segmentation, outputting the final PCI planning scheme. By globally verifying and adjusting interference across subgraph boundaries, the potential defects of decomposition optimization are compensated for, ensuring that locally optimal schemes can be integrated into globally consistent, high-quality plans.

[0013] In a second aspect, this invention provides a heuristic-based cell PCI planning system for executing the method, comprising: a model building module, configured to construct a multi-objective optimization model based on the needs of PCI planning in a communication network, with the objective of minimizing the number of collision MRs, confusion MRs, and modulo 3 interference MRs in the communication network, and at least with PCI allocation uniqueness constraints, collision avoidance constraints, confusion avoidance constraints, and modulo 3 interference avoidance constraints as constraints; a data acquisition and initialization module, configured to read data of each cell and its neighboring relationships, and initialize the PCI of all cells to an unallocated state; an initial allocation module, configured to use data of each cell and its neighboring relationships as input to the multi-objective optimization model, and perform initial allocation using a heuristic algorithm to generate an initial PCI allocation scheme that satisfies the constraints; a problem decomposition module, configured to construct a cell interaction graph model based on data of each cell and its neighboring relationships, and use a graph partitioning algorithm to partition the cell interaction graph model so that the multi-objective optimization problem is decomposed into multiple sub-optimization problems; and a parallel optimization module, configured to solve the multi-objective optimization model in parallel using a multi-objective evolutionary algorithm for each sub-optimization problem, in conjunction with the initial PCI allocation scheme, to obtain a set of optimal PCI allocation schemes.

[0014] Compared with the prior art, the present invention has the following advantages: First, a multi-objective optimization model is constructed with the goals of minimizing the conflict MR number, confusion MR number, and modulo 3 interference MR number. This model fully considers the conflict MR number, confusion MR number, and modulo 3 interference MR number, thus providing precise optimization guidance and feasibility criteria for the entire solution. Next, a heuristic algorithm is used to quickly generate an initial feasible solution that satisfies all constraints, effectively narrowing the search range for high-quality solutions and providing a good starting point for subsequent complex optimizations, thereby improving solution efficiency. Then, a cell interaction graph model is constructed based on the data of each cell and its neighboring relationships. This model abstracts network entities and interference relationships into an interaction graph model, fully considering the interference relationships between neighboring cells. A graph partitioning algorithm is used to intelligently divide this into multiple internally related subgraphs, thus decomposing the large-scale, globally strongly coupled complex optimization problem into multiple small-scale, parallelizable sub-optimization problems. This not only effectively solves the conflict and interference problems between neighboring cells but also overcomes the complexity challenge of directly solving the model in large-scale networks. Finally, for each subproblem, a multi-objective evolutionary algorithm is run in parallel. Guided by the aforementioned initial solution, the algorithm efficiently searches the solution space of each subproblem for the optimal set of solutions that can balance the three objectives of conflict, confusion and modulo 3 interference, thereby achieving the coordinated optimization of the conflict MR number, confusion MR number and modulo 3 interference MR number.

[0015] In summary, this invention, by constructing an accurate multi-objective optimization model and designing an efficient decomposition-parallel system solution framework, ensures that while guaranteeing solution efficiency and handling ultra-large-scale networks, the solution can systematically minimize the three major interference indicators at each local level and even globally (globally optimizing conflicts, confusion, and modulo 3 interference while ensuring efficiency), ultimately outputting a high-quality, feasible global PCI planning scheme. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 This is a flowchart of the heuristic cell PCI planning method provided in Embodiment 1 of the present invention; Figure 2 This is a block diagram of the heuristic-based cell PCI planning system provided in Embodiment 1 of the present invention; Figure 3 This is a flowchart of the heuristic cell PCI planning method provided in Embodiment 2 of the present invention; Figure 4This is a block diagram of the heuristic-based cell PCI planning system provided in Embodiment 2 of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: Please refer to Figure 1 This embodiment provides a heuristic-based cell PCI planning method, mainly implemented through steps S0 to S4. The specific implementation of each step will be described in detail below: S0. Based on the requirements of PCI planning in the communication network, a multi-objective optimization model is constructed with the goal of minimizing the number of conflict MRs, confusion MRs, and modulo 3 interference MRs in the communication network, and with at least the PCI allocation uniqueness constraint, conflict avoidance constraint, confusion avoidance constraint, and modulo 3 interference avoidance constraint as constraints.

[0020] First, we determine the objective functions as follows: (1) Minimize the number of collision MRs between cells. The objective function is: ; in, It represents the number of conflicting MRs between cells i and j; is a decision variable, taking the value 0 or 1. When it is 1, it means that cells i and j are assigned the same PCI, otherwise it is 0; N is the maximum capacity of the cell.

[0021] (2) Minimize the number of confusion MRs between cells. The objective function is: ; in, It is the number of confusion MRs between cells i and j; is a decision variable, taking the value 0 or 1. When it is 1, it means that cells i and j are assigned the same PCI, otherwise it is 0; N is the maximum capacity of the cell.

[0022] (3) Minimize the number of inter-cell interference MRs, with the objective function being: ; in, It is the number of Mode 3 interference (MR) between cells i and j; This is a decision variable, taking the value 0 or 1. When it is 1, it means that cells i and j are assigned the same mod 3 value; otherwise, it is 0.

[0023] Secondly, set the constraints as follows: (1) PCI allocation uniqueness constraint: The PCI allocation uniqueness constraint means that each cell is assigned a unique PCI value. In this embodiment, k represents the PCI values ​​of cells i and j, and the PCI is used to classify and label the cells.

[0024] Let the variable Let k be a binary function, and let the PCI value of cell i be k. Otherwise, if the PCI value of cell i is not k, then Considering that each cell has one and only one PCI value, The PCI assignment uniqueness constraint parameter is expressed as follows: ; (2) Conflict avoidance constraint and confusion avoidance constraint. Conflict avoidance constraint means avoiding assigning the same PCI to any two cells with the same frequency; confusion avoidance constraint means avoiding assigning the same PCI to any two cells with the same frequency and that are neighboring cells.

[0025] When cells i and j have the same frequency and are assigned the same PCI... The expression is: 1 if the value is 1, otherwise 0. ; Considering that nonlinear piecewise functions increase the difficulty of solving the problem, and Convert to: ; ; Where N is the maximum cell capacity, q represents the maximum number of cells i and j that are on the same frequency and assigned the same PCI value, h represents the number of cells i and j that are on the same frequency and assigned the same PCI value, and n is the sequence of two constraints to avoid conflict and confusion.

[0026] (3) Modulo 3 interference avoidance constraint, which means avoiding assigning PCI with the same modulo 3 value to any two cells.

[0027] To determine whether the PCI values ​​of cells i and j have the same remainder after being divided by 3, The value of PIC modulo 3 is 0 if cells i and j have the same value, indicating interference; otherwise, it is 0. This can be expressed mathematically as: ; Where k and k' represent the PCI values ​​allocated to cells i and j, respectively.

[0028] (4) Binary variable constraints: In this embodiment, all decision variables are binary variables, and the constraints are as follows: ; ; ; Then, the multi-objective optimization model constructed using the above objective function and constraints is solved. The solution process is implemented through S1 to S4: S1. Read the data of each cell and its neighbor relationships, and initialize the PCI of all cells to an unallocated state.

[0029] S2. Using the data of each cell and its neighbor relationships as input to the multi-objective optimization model, an initial allocation is performed using a heuristic algorithm to generate an initial PCI allocation scheme that satisfies the constraints. In a preferred embodiment, in step S2, the heuristic algorithm uses a greedy algorithm, and the initial allocation process includes: calculating the adjacency degree of each cell based on the neighbor relationships; sorting the cells according to their adjacency degrees, starting with the cell with the highest adjacency degree, and allocating PCIs to each cell in descending order of adjacency degree, ensuring that the PCIs of each cell are different from those of all its neighboring cells.

[0030] S3. Construct a cell interaction graph model based on the data of each cell and its neighboring relationships, and use a graph segmentation algorithm to segment the cell interaction graph model so that the multi-objective optimization problem is decomposed into multiple sub-optimization problems.

[0031] In a preferred embodiment, step S3 involves constructing a cell interaction graph model based on the data of each cell and its neighbor relationships. Specifically, this includes: treating each cell as a node in the graph; establishing an edge between any two adjacent cells and their corresponding nodes based on the input neighbor relationship data; and assigning corresponding weights to this edge based on the input collision MR number, confusion MR number, and modulo 3 interference MR number, thereby forming a weighted cell interaction graph model. A more detailed implementation process is as follows: (1) Define nodes: Each cell to be planned is abstracted as a node in the graph model. For example, if there are 1000 cells in the network, there will be 1000 nodes in the graph; (2) Define edges (topology): Based on the proximity data between cells input in step S1, if the data indicates that cell A and cell B are adjacent, then an undirected edge is established between the two nodes representing A and B. In this way, all the proximity relationships of all nodes constitute the edge set of the graph, which describes the connection topology of the network; (3) Define edge weights (quantify interference intensity): This step is crucial in transforming the raw data into an optimized guided model. Assign one or more weight values ​​to each edge. These weights are directly derived from another part of the S1 input data: conflict MR count, confusion MR count, and modulo 3 interference MR count. For example, a comprehensive weight W_AB can be set for the edge connecting node A and node B, calculated as: W_AB = α (Conflict MR number_AB)+β (confusion MR number _AB)+γ (Modal 3 Interference MR Number_AB); where α, β, and γ are preset coefficients used to balance the importance of the three types of interference. The higher the value of the weight W_AB, the more severe the mutual interference between the two cells, and the more priority it needs to be given in optimization.

[0032] The completed cell interaction graph model: Finally, through the above steps, a standard weighted undirected graph G=(V,E,W) will be obtained: V (set of nodes): all cells. E (set of edges): all proximity relationships. W (set of weights): the quantized value of the interference intensity corresponding to each edge (based on MR data).

[0033] In a preferred embodiment, to handle the mutual influence between large-scale cells, the Louvain algorithm is used to decompose the large-scale cell interaction model into multiple smaller-scale sub-problems. Specifically, in step S3, the graph segmentation algorithm employs the Louvain algorithm, which automatically segments the cell interaction graph model into multiple internally connected subgraphs by iteratively optimizing the modularity index, thereby decomposing the multi-objective optimization problem into sub-optimization problems corresponding to each subgraph. It is understood that the Louvain algorithm is a classic community detection algorithm, also known as a graph segmentation algorithm, that identifies tightly connected subgroups (i.e., "communities" or "subgraphs") from complex networks. Its core objective is to maximize modularity—an index that measures the quality of network community partitioning.

[0034] Modularity is a quantitative metric for measuring the quality of a subgraph partition. Its value ranges from -1 to 1. A higher value (closer to 1) indicates a better partition. That is, the connections within a subgraph are very tight (strong inter-subgraph interference), while the connections between different subgraphs are very sparse (weak cross-subgraph interference). This is the ideal state to achieve when decomposing problems. A lower value indicates a random or poor partition, with no significant difference between internal and external connections within the subgraph. Iterative optimization describes how the Louvain algorithm works: it repeatedly tries moving nodes to different subgraphs to observe whether it improves the overall modularity value. This process is repeated until the modularity can no longer be improved, at which point the (locally) optimal partitioning scheme is obtained. In short, the Louvain algorithm uses an iterative, step-by-step improvement approach to find the best grouping scheme that results in "strong clustering within each subgraph and loose connections between subgraphs." In the PCI planning method of this embodiment, the direct purpose of optimizing the modularity index is to automatically generate high-quality "subgraph" partitions, thereby ensuring high independence of subproblems (within the subgraph). Due to the sparse connections between subgraphs, optimizing the PCI allocation scheme within one subgraph has minimal impact on other subgraphs. This allows subsequent parallel optimization (step S4) to be performed approximately independently. It avoids the algorithm generating meaningless, random partitions, ensuring that each subgraph is a "highly cohesive, loosely coupled" optimization unit.

[0035] S4. For each sub-optimization problem, the multi-objective optimization model is solved in parallel using a multi-objective evolutionary algorithm in conjunction with the initial PCI allocation scheme to obtain the set of optimal PCI allocation schemes.

[0036] In a preferred embodiment, in step S4, the multi-objective evolutionary algorithm employs the NSGA-II algorithm. The solution process includes: initializing the population under the premise of satisfying constraints; evaluating individual fitness and performing non-dominated sorting according to the objective function in the multi-objective optimization model; calculating the crowding distance of individuals to maintain population diversity; generating offspring populations through selection, crossover, and mutation operations; merging parent and offspring populations and selecting a new generation population; iterative evolution until termination. Specifically, during population initialization, the PCI allocation schemes corresponding to the subgraphs in the initial PCI allocation scheme are included as a complete individual in the initial population. In a preferred embodiment, the step of evaluating individual fitness and performing non-dominated sorting according to the objective function in the multi-objective optimization model specifically evaluates fitness by determining the Pareto rank of each individual. In step S4, the obtained set of optimal PCI allocation schemes is the set of Pareto-optimal PCI allocation schemes in each subgraph. In a preferred embodiment, in the step of generating the offspring population through selection, crossover, and mutation operations, the selection operation is specifically based on a tournament selection mechanism, selecting individuals with fitness and crowding exceeding a preset threshold for subsequent crossover and mutation operations. In this embodiment, based on graph partitioning, NSGA-II is applied to optimize each subproblem; the more specific solution process is as follows: a. Initialize the population: Randomly generate an initial population, with each individual representing a PCI allocation scheme; b. Non-dominated ranking: Perform non-dominated ranking on the population to determine the Pareto rank of each individual; c. Crowding Calculation: Calculate the crowding distance for each individual to maintain population diversity; d. Selection: Based on the tournament selection mechanism, individuals with high fitness and high crowding are selected for crossover and mutation operations; e. Crossover and mutation: Perform crossover and mutation operations to generate new offspring populations; f. Merging populations: Merging the parent population and the offspring population to form a new population; g. Selecting a new generation of population: Select the top N best individuals from the merged population to form a new generation of population; h. Iteration: Repeat the above steps until the predetermined number of iterations is reached or the convergence condition is met.

[0037] Please refer to Figure 2Corresponding to the heuristic-based cell PCI planning method of this embodiment, a heuristic-based cell PCI planning system is provided to execute the method of this embodiment. It mainly consists of the following modules: a model building module 0, used to construct a multi-objective optimization model based on the PCI planning requirements in the communication network, aiming to minimize the number of collision MRs, confusion MRs, and modulo 3 interference MRs in the communication network, and using at least PCI allocation uniqueness constraints, collision avoidance constraints, confusion avoidance constraints, and modulo 3 interference avoidance constraints as constraints; and a data acquisition and initialization module 1, used to read the data of each cell and its neighbor relationships, and initialize the PCI of all cells. Unassigned state; Initial allocation module 2, used as input to the multi-objective optimization model with data on each cell and its neighboring relationships, employs a heuristic algorithm for initial allocation, generating an initial PCI allocation scheme that satisfies the constraints; Problem decomposition module 3, used to construct a cell interaction graph model based on data on each cell and its neighboring relationships, and uses a graph partitioning algorithm to partition the cell interaction graph model, so that the multi-objective optimization problem is decomposed into multiple sub-optimization problems; Parallel optimization module 4, used to solve the multi-objective optimization model in parallel using a multi-objective evolutionary algorithm for each sub-optimization problem, combined with the initial PCI allocation scheme, to obtain the optimal set of PCI allocation schemes.

[0038] Example 2: Please refer to Figure 3 This embodiment provides a more preferred heuristic-based cell PCI planning method. The specific implementation steps are based on steps S0 to S4 given in embodiment 1, with the addition of step S5.

[0039] S5. Merge the set of optimal PCI allocation schemes to generate candidate global schemes, and perform conflict and interference checks and adjustments on cells with proximity relationships across subgraph boundaries generated by subgraph segmentation, outputting the final PCI planning scheme. Specifically, the implementation process is as follows.

[0040] First, the process of fusing and generating candidate solutions: Since S4 outputs a Pareto optimal solution set (i.e., multiple sets of high-quality solutions) for each subgraph, S5 needs to select one solution from the solution set of each subgraph and combine them to form a complete candidate global solution. The selection strategy can be a simple random selection or an optimal combination selection based on a fast global cost evaluation of the solutions in each subgraph.

[0041] Next, targeted verification: The algorithm locates and locks those connections that were "cut off" during the S3 graph segmentation—that is, "cross-subgraph boundary cell pairs." It systematically traverses all these specific boundary cell pairs, checking them one by one against the constraints (conflict, confusion, modulo 3 interference) in the current candidate global scheme for violations of PCI allocation, based on the constraints in the multi-objective optimization model established in S0.

[0042] Finally, intelligent adjustment and output: If a violation is detected during verification (e.g., two boundary cells that should avoid conflict are assigned the same PCI), a local adjustment procedure is initiated. The adjustment, without disrupting the optimized PCI allocation structure within each subgraph, prioritizes minimal disturbance. For example, it reallocates PCI resources only within the available PCI resource pool of the one or a few boundary cells involved in the conflict, thus eliminating the conflict. This process may iterate until all cross-boundary conflicts and interferences are eliminated, ultimately outputting a final PCI planning scheme that satisfies all constraints at the global level and is composed of high-quality solutions from each subgraph.

[0043] Please refer to Figure 4 Corresponding to the heuristic-based cell PCI planning method of this embodiment, a heuristic-based cell PCI planning system is provided to execute the method of this embodiment. Based on the model building module 0, data acquisition and initialization module 1, initial allocation module 2, problem decomposition module 3, and parallel optimization module 4 given in embodiment 1, a fusion and adjustment module 5 is added to fuse the set of optimal PCI allocation schemes, generate candidate global schemes, and perform conflict and interference verification and adjustment on cells with proximity relationships across subgraph boundaries generated by subgraph segmentation, and output the final PCI planning scheme.

[0044] This invention, through the above embodiments, addresses the large-scale, complex PCI configuration and planning problem by combining greedy algorithms, graph partitioning algorithms, and the NSGA-II algorithm to achieve efficient and near-globally optimal cell PCI allocation and planning. Simultaneously, by fully considering the complex interference relationships between cells, it effectively resolves conflicts and interference issues between neighboring cells. Especially in large-scale, high-density network environments, this invention's method can achieve global optimization, minimizing the number of conflict MR (Measurement Report) counts, confusion MR counts, and modulo 3 interference MR counts in the network, thereby effectively improving key performance indicators such as user rate and handover success rate.

[0045] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0046] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification. Furthermore, the above embodiments only illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. For those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all fall within the protection scope of the present invention.

Claims

1. A heuristic-based cell PCI planning method, characterized in that, Includes the following steps: S0. Based on the requirements of PCI planning in the communication network, a multi-objective optimization model is constructed with the goal of minimizing the number of collision MRs, confusion MRs, and modulo 3 interference MRs in the communication network, and with at least the PCI allocation uniqueness constraint, collision avoidance constraint, confusion avoidance constraint, and modulo 3 interference avoidance constraint as constraints. S1. Read the data of each cell and its neighbor relationships, and initialize the PCI of all cells to an unallocated state; S2. Using the data of each cell and its neighbor relationships as input to the multi-objective optimization model, a heuristic algorithm is used for initial allocation to generate an initial PCI allocation scheme that satisfies the constraints. S3. Construct a cell interaction graph model based on the data of each cell and its neighboring relationships, and use a graph segmentation algorithm to segment the cell interaction graph model so that the multi-objective optimization problem is decomposed into multiple sub-optimization problems; S4. For each sub-optimization problem, the multi-objective optimization model is solved in parallel using a multi-objective evolutionary algorithm in conjunction with the initial PCI allocation scheme to obtain the set of optimal PCI allocation schemes.

2. The heuristic-based cell PCI planning method according to claim 1, characterized in that, In step S0, the constraints also include binary variable constraints; the PCI allocation uniqueness constraint means that each cell is assigned a unique PCI value; the conflict avoidance constraint means that the same PCI is avoided for any two co-frequency cells; the confusion avoidance constraint means that the same PCI is avoided for any two co-frequency and neighboring cells; and the modulo-3 interference avoidance constraint means that the same modulo-3 value is avoided for any two cells.

3. The heuristic-based cell PCI planning method according to claim 1, characterized in that, In step S2, the heuristic algorithm adopts a greedy algorithm. The initial allocation process includes: calculating the adjacency degree of each cell based on the proximity relationship; sorting the cells according to their adjacency degree, starting the allocation from the cell with the highest adjacency degree, and allocating PCI to each cell in descending order of adjacency degree to ensure that the PCI of each cell is different from that of all its neighboring cells.

4. The heuristic-based cell PCI planning method according to claim 1, characterized in that, In step S3, a cell interaction graph model is constructed based on the data of each cell and its neighbor relationships. Specifically, this includes: taking each cell as a node in the graph; establishing an edge between any two nodes corresponding to cells that have a neighbor relationship based on the input neighbor relationship data; and assigning corresponding weights to the edge based on the input conflict MR number, confusion MR number and modulo 3 interference MR number, thereby forming a weighted cell interaction graph model.

5. The heuristic-based cell PCI planning method according to claim 1, characterized in that, In step S3, the graph segmentation algorithm specifically adopts the Louvain algorithm, which automatically segments the cell interaction graph model into multiple internally connected subgraphs by iteratively optimizing the modularity index, so as to decompose the multi-objective optimization problem into sub-optimization problems corresponding to each subgraph.

6. The heuristic-based cell PCI planning method according to claim 5, characterized in that, In step S4, the multi-objective evolutionary algorithm adopts the NSGA-II algorithm, and the solution process includes: initializing the population under the premise of satisfying the constraints; evaluating the fitness of individuals according to the objective function in the multi-objective optimization model and performing non-dominated sorting; calculating the crowding distance of individuals to maintain population diversity; generating offspring populations through selection, crossover and mutation operations; merging the parent and offspring populations and selecting to obtain a new generation population; iterative evolution until termination; wherein, when initializing the population, the PCI allocation scheme of the corresponding subgraph in the initial PCI allocation scheme is included as a complete individual in the initial population.

7. The heuristic-based cell PCI planning method according to claim 6, characterized in that, In the step of evaluating individual fitness and performing non-dominated ranking based on the objective function in the multi-objective optimization model, fitness is specifically evaluated by determining the Pareto level of each individual; in step S4, the set of optimal PCI allocation schemes obtained is: the set of Pareto optimal PCI allocation schemes in each subgraph.

8. The heuristic-based cell PCI planning method according to claim 6, characterized in that, In the step of generating offspring population through selection, crossover, and mutation operations, the selection operation is specifically based on the tournament selection mechanism, selecting individuals with fitness and crowding higher than a preset threshold for subsequent crossover and mutation operations.

9. The heuristic-based cell PCI planning method according to claim 1, characterized in that, Also includes: S5. Integrate the set of optimal PCI allocation schemes to generate candidate global schemes, and perform conflict and interference verification and adjustment on cells with proximity relationships across subgraph boundaries generated by subgraph segmentation, and output the final PCI planning scheme.

10. A heuristic-based cell PCI planning system for performing the method described in any one of claims 1 to 9, characterized in that, include: The model building module is used to construct a multi-objective optimization model based on the needs of PCI planning in the communication network, with the goal of minimizing the number of collision MRs, confusion MRs and modulo 3 interference MRs in the communication network, and with at least PCI allocation uniqueness constraints, collision avoidance constraints, confusion avoidance constraints and modulo 3 interference avoidance constraints as constraints. The data acquisition and initialization module is used to read data on each cell and its neighboring relationships, and initialize the PCI of all cells to an unallocated state; The initial allocation module is used to take the data of each cell and its neighbor relationships as input to the multi-objective optimization model, and to perform initial allocation using a heuristic algorithm to generate an initial PCI allocation scheme that satisfies the constraints. The problem decomposition module is used to construct a cell interaction graph model based on data of each cell and its neighboring relationships, and to use a graph segmentation algorithm to segment the cell interaction graph model so that the multi-objective optimization problem is decomposed into multiple sub-optimization problems; The parallel optimization module is used to solve the multi-objective optimization model in parallel using a multi-objective evolutionary algorithm for each sub-optimization problem, in conjunction with the initial PCI allocation scheme, to obtain the set of optimal PCI allocation schemes.

Images