A co-60 radiation source rod new scene arrangement optimization method based on DCM and IASA algorithm

Abstract

This invention discloses a method for optimizing the arrangement of Co-60 radiation source rods in a newly constructed scene based on DCM and IASA algorithms, belonging to the technical field of radiation source rod arrangement, including: constructing Co... 60 A basic radiation field model for a newly constructed scenario using a radiation source rod; a pre-calculated dose coefficient matrix is ​​constructed based on the basic radiation field model, and this pre-calculated dose coefficient matrix is ​​used as Co. 60 A dose assessment acceleration operator in the iterative process of radiation source rod arrangement optimization; an improved adaptive simulated annealing algorithm is used for Co. 60 The radiation source rod arrangement scheme is subjected to global search and iterative optimization, and the final output is the Co value corresponding to the minimum dose inhomogeneity of the irradiation field. 60 The globally optimal arrangement of radiation source rods. This invention achieves significant improvements in computational efficiency and optimization quality through deep synergy between low-level computational paradigm acceleration and high-level search strategy optimization, providing a basis for the construction of new Co... 60 The intelligent and sophisticated design of irradiation facilities provides strong and reliable technical support.

Description

Technical Field

[0001] This invention belongs to the technical field of radiation source rod arrangement, specifically relating to a method for optimizing the arrangement of Co-60 radiation source rods in a newly created scene based on DCM and IASA algorithms. Background Technology

[0002] Irradiation processing technology is an important component of civilian nuclear technology. Its working principle involves inducing chemical or biological changes in irradiated materials through ionizing radiation, thereby achieving purposes such as material modification, medical sterilization, and food preservation. 60 Gamma irradiation technology, due to its high penetration and stable operation, has become the cornerstone of the modern irradiation processing industry. In irradiation station design, the uniformity of dose distribution within the radiation field is a key indicator determining processing quality and efficiency, and the spatial arrangement of the radiation source rods on the source frame plane is a crucial determinant of dose uniformity. Therefore, intelligently and efficiently optimizing the arrangement of radiation source rods to minimize the non-uniformity of the overall dose field is a core engineering challenge.

[0003] Currently, some research has been conducted on optimizing the arrangement of source rods in newly constructed irradiation facilities, with mainstream optimization methods primarily relying on various heuristic optimization algorithms. For example, plant growth simulation algorithms search for the optimal layout by simulating the phototropic growth mechanism of plants; genetic algorithms iteratively improve source rod configurations by drawing on selection, crossover, and mutation operations from biological evolution. These methods have, to some extent, achieved exploration of high-dimensional, non-convex solution spaces. However, existing technologies have revealed significant limitations in practical industrial applications, mainly in the following two aspects: 1. The computational efficiency bottleneck is prominent. Existing mainstream optimization algorithms, such as plant growth simulation algorithms and genetic algorithms, exhibit computational complexity that increases superlinearly or even factorially with the number of source rods. For a typical industrial-scale irradiation station with approximately 80 source rods, the solution space is about 80!, and the computational cost far exceeds the actual processing capacity of traditional algorithms. This results in extremely long optimization processes, failing to meet the needs of rapid iteration and real-time decision-making in engineering design, and severely restricting the application of optimization methods in the design phase of large-scale new projects.

[0004] 2. Lack of effective computational acceleration mechanisms. During the optimization iteration process, each evaluation of the adaptability of candidate layouts, i.e., dose non-uniformity, requires recalculating the dose contribution of all source rods to each reference point. Traditional methods employ real-time geometric calculations, with a computational complexity of O(U*V) (U being the number of source rods and V being the number of reference points). When the number of source rods and reference points is large, the computational overhead of a single dose evaluation is enormous, becoming the main performance bottleneck of the entire optimization process. Although the algorithm itself is continuously being improved, the lack of critical optimization for the core time-consuming underlying dose calculation results in limited overall efficiency improvement.

[0005] In summary, existing technologies suffer from two main drawbacks: excessively long optimization time and significant computational efficiency bottlenecks. This makes it difficult for current methods to optimize large-scale new Co-type structures within an acceptable timeframe. 60 Irradiation facilities provide high-quality, highly uniform source rod layout schemes. Therefore, there is an urgent need for a new solution that integrates efficient optimization algorithms with underlying computational acceleration technologies to improve computational efficiency while ensuring optimization quality. Summary of the Invention

[0006] The purpose of this invention is to address the aforementioned shortcomings in the prior art by providing a method for optimizing the arrangement of Co-60 radiation source rods in a newly constructed scene based on the DCM and IASA algorithms, thereby solving the aforementioned technical problems. Existing Co 60 The optimization method for source rod arrangement in irradiation facilities suffers from excessively high computational complexity in dose assessment (O(U*V)) and slow convergence speed. This results in excessively long computation time for optimization tasks in typical industrial-scale irradiation stations (e.g., 80 source rods), failing to meet the timeliness requirements of engineering design and becoming a major efficiency bottleneck restricting the practical application of the technology. At the same time, traditional heuristic algorithms are prone to getting trapped in local optima in high-dimensional solution spaces, and the optimization results are highly sensitive to initial conditions and parameter settings. It is difficult to maintain the stability of the process and the robustness of the algorithm while ensuring high optimization quality (low dose non-uniformity), thus affecting the dose uniformity and consistency of the final layout scheme and reducing the reliability and repeatability of irradiation processing quality.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for optimizing the arrangement of Co-60 radiation source rods in a newly constructed scene based on DCM and IASA algorithms includes the following steps: S1, Construct Co 60 The basic model of the radiation field in the newly constructed scene of the radiation source rod; S2. Construct a pre-calculated dose coefficient matrix based on the aforementioned radiation field fundamental model, and use the pre-calculated dose coefficient matrix as Co. 60 A dose assessment acceleration operator in the iterative process of radiation source rod arrangement optimization; S3. Based on the aforementioned dose assessment acceleration operator, an improved adaptive simulated annealing algorithm is used to process Co. 60 The radiation source rod arrangement scheme is subjected to global search and iterative optimization, and the final output is the Co value corresponding to the minimum dose inhomogeneity of the irradiation field. 60 The globally optimal arrangement of radiation source rods.

[0008] Furthermore, step S1 includes the following sub-steps: S11. Define a radiation source plane in the XOZ coordinate system plane, and set a source slot in the radiation source plane to represent the placement position of the source rod. The coordinate set of the source slot is: In the formula, This represents the set of coordinates of the source slot within the radiation source plane; This represents the coordinates of the vertex on the source slot. Let x be the x-coordinate of the vertex on the source slot. Let be the vertical coordinate of the vertex on the source slot. , These represent the number of rows and columns of the source slot, respectively. and These represent the total number of rows and columns of the source slots, respectively. S12. Define a dose reference plane parallel to the radiation source plane. The dose reference plane is located at the vertical coordinate Y=100cm, and its coordinate set is as follows: In the formula, The set of coordinates representing the reference plane; Indicates the coordinates of the reference point. The x-coordinate of the reference point The vertical coordinate of the reference point, The row number of the reference point. The number of columns for the reference point. , These represent the total number of rows and columns of the reference point, respectively. S13. Based on the radiation source plane and dose reference plane, construct Co 60 A three-dimensional parametric geometric model of an irradiation station; S14. In a three-dimensional parametric geometric model, calculate a single root Co. 60 The radiation dose of the radiation source rod to the irradiation reference point O; S15. Based on the radiation dose calculation results in S14, calculate all Co... 60 The total dose of the radiation source rod to a single reference point is calculated, and the dose non-uniformity of the reference plane is calculated. S16. In the newly constructed irradiation station scenario, based on dose non-uniformity, Co... 60 The task of optimizing the arrangement of radiation source rods is formalized as a high-dimensional non-convex programming problem.

[0009] Furthermore, in S14, the calculation of a single Co root 60 The radiation dose from the radiation source rod to the irradiation reference point O is expressed as: In the formula, For the first Line 1 The radiation dose of the source rod to the irradiation reference point O, The irradiation constant, For source rod activity, This represents the perpendicular distance between any irradiated point in the reference plane and the source rod. Indicates the length of the source rod. This represents the distance from the foot of the perpendicular to the lower vertex of the source rod.

[0010] Furthermore, in S15, all Co are calculated. 60 The total dose from the radiation source rod to a single reference point is expressed as: In the formula, Total dose; Based on the total dose, the dose non-uniformity of the reference plane is calculated, and it is expressed as: In the formula, Dose nonuniformity at the reference plane.

[0011] Furthermore, in S16, Co... 60 The task of optimizing the arrangement of radiation source rods is formalized as a high-dimensional non-convex programming problem, and the search is performed to find the Co that minimizes the dose non-uniformity of the reference plane. 60 The arrangement and combination of radiation source rods is the optimization objective, which is expressed as: In the formula, This represents the minimum dose nonuniformity obtained under the optimal source rod arrangement scheme.

[0012] Furthermore, step S2 includes the following sub-steps: S21, The pre-calculated dose coefficient matrix is ​​constructed by: taking the irradiation dose in S14... The expression is separated into the geometric influence factor and the source rod activity A, and then the dose contribution matrix B is constructed. The total dose distribution D is then transformed into a matrix multiplication operation between the source rod activity A and the dose contribution matrix B, i.e.: ; S22, Dose Contribution Matrix B in Co 60 The radiation source rod arrangement optimization is calculated and stored once before the iteration, and can be repeatedly called in subsequent optimization iterations without recalculating the geometric influence factor, thus eliminating redundant calculation overhead.

[0013] Furthermore, in S21, the geometric influence factor is expressed as: In the formula, For dose contribution matrix The element in represents the slot number. OK The geometric influence factor of the source bar at column n on the reference point plane at row m.

[0014] Furthermore, step S3 includes the following sub-steps: S31. Based on the source-slot coordinate parameters specified by the user, establish the initial spatial configuration of the irradiation system that matches the basic model of the radiation field; S32. Based on the dose reference plane defined in S12, set several reference points for dose assessment; S33, according to Co 60 Given the known activity distribution of the radiation source rods, an initial arrangement scheme of the source rods is generated, and the dose assessment acceleration operator is called to calculate the dose non-uniformity corresponding to the initial arrangement scheme. S34. Using the initial source rod arrangement scheme as the initial solution for iteration, a new source rod arrangement scheme is generated through a neighborhood search strategy, and the dose evaluation acceleration operator is called to calculate the dose non-uniformity of the new arrangement scheme. S35. Determine whether to accept the new source bar arrangement scheme according to the Metropolis criterion. If the new solution is better, accept it directly; otherwise, accept the suboptimal solution with a preset probability. S36. Repeat S34~S35 for multiple rounds of iterative optimization until the algorithm converges, and output Co. 60 The globally optimal arrangement of radiation source rods.

[0015] Furthermore, in S35, the Metropolis criterion is: In the formula, This represents the probability of accepting the new solution. and These represent the current solution and the optimized new solution, respectively. This represents the temperature parameter in the simulated annealing algorithm.

[0016] The method for optimizing the arrangement of Co-60 radiation source rods in a newly constructed scene based on DCM and IASA algorithms provided by this invention has the following beneficial effects: This invention, by constructing a technical framework for the collaborative optimization of a "pre-calculated dose coefficient matrix (DCM)" and an "improved adaptive simulated annealing algorithm (IASA)," produces the following significant beneficial effects compared to existing technologies. These effects directly stem from and closely correspond to the aforementioned key technical solutions: 1. A breakthrough in computational efficiency was achieved, solving the timeliness problem of large-scale optimization; Existing technologies require real-time calculation of the complex geometric dose contribution between all source rods and the reference point in each iteration, resulting in a computational complexity of O(U*V), which has become a major performance bottleneck. This invention introduces a core accelerator, the "pre-calculated dose coefficient matrix (DCM)," which calculates and stores all fixed factors related to geometric location in a single operation before optimization. During optimization iterations, dose assessment is transformed into a single efficient matrix multiplication operation (D=A·B), reducing the complexity to O(V). According to experimental data, this innovation reduces runtime from approximately 247.6 seconds to approximately 0.47 seconds at the same optimization quality, a speedup of nearly 500 times. This makes rapid, multi-stage optimization of newly built industrial facilities with 80 or more source rods a reality, meeting the engineering design requirements for real-time or near-real-time response.

[0017] 2. Existing heuristic algorithms are prone to getting trapped in local optima and exhibit significant fluctuations in results. The "Improved Adaptive Simulated Annealing (IASA)" proposed in this invention effectively enhances global exploration capabilities and local development accuracy in high-dimensional solution spaces by combining dynamic neighborhood search strategies (exchange, reversal, etc.) and an adaptive annealing schedule. The rapid evaluation provided by DCM allows the algorithm to perform far more iterations than traditional methods in the same amount of time, thus more thoroughly searching the solution space. As shown in Tables 3 and 4, the IASA algorithm achieves a significantly better average dose nonuniformity (μ=1.06) in ten independent runs than several mainstream optimization algorithms compared to it (such as DRA's 1.17, HGS's 1.14, etc.), with a low standard deviation, demonstrating excellent convergence accuracy and stability, and providing a better and more reliable source rod layout scheme for facilities.

[0018] In summary, this invention achieves significant improvements in computational efficiency and optimization quality through deep synergy between low-level computational paradigm acceleration (DCM) and high-level search strategy optimization (IASA), providing a foundation for the creation of new Co... 60 The intelligent and sophisticated design of irradiation facilities provides strong and reliable technical support. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method for optimizing the arrangement of Co-60 radiation source rods in a new scene based on the DCM and IASA algorithms in the embodiment.

[0020] Figure 2 This is a schematic diagram of the three-dimensional coordinate system of the source plane and the reference plane in the embodiment.

[0021] Figure 3 This is a schematic diagram of the three-dimensional coordinates of the source rod and the reference point in the embodiment.

[0022] Figure 4 The flowchart below shows the improved simulated annealing algorithm in this embodiment. Figure 5 This is a schematic diagram of the source frame modeling in the embodiment.

[0023] Figure 6 This is a comparison chart of algorithm non-uniformity in the embodiments. Detailed Implementation

[0024] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0025] This embodiment provides a method for optimizing the arrangement of Co-60 radiation source rods in a newly constructed scenario based on the Dose Coefficient Matrix (DCM) and IASA algorithms. Its core lies in constructing an "algorithm-accelerator" collaborative optimization framework. It achieves orders-of-magnitude acceleration of underlying dose calculations through pre-computed Dose Coefficient Matrix (DCM) and utilizes an improved Adaptive Simulated Annealing (IASA) algorithm for efficient and robust global search in a high-dimensional solution space, ultimately outputting the optimal or near-optimal source rod spatial arrangement scheme. (Refer to...) Figure 1 Specifically, it includes the following: S1, Construct Co 60 The basic model of the radiation field for the newly created scene of the radiation source rod; including the following sub-steps: S11. Source plane definition: Define a radiation source plane in the XOZ coordinate system, and set a slot (referred to as a source slot) in the radiation source plane to characterize the placement position of the source rod. The coordinate set of the source slot is: In the formula, This represents the set of coordinates of the source slot within the radiation source plane; This represents the coordinates of the vertex on the source slot. Let x be the x-coordinate of the vertex on the source slot. Let be the vertical coordinate of the vertex on the source slot. , These represent the number of rows and columns of the source slot, respectively. and These represent the total number of rows and columns of the source slots, respectively. S12. Definition of Dose Reference Plane; Define a dose reference plane parallel to the radiation source plane, wherein the dose reference plane is located at the vertical coordinate Y=100cm, and its coordinate set is: In the formula, The set of coordinates representing the reference plane; Indicates the coordinates of the reference point. The x-coordinate of the reference point The vertical coordinate of the reference point, The row number of the reference point. The number of columns for the reference point. , These represent the total number of rows and columns of the reference point, respectively.

[0026] In one specific embodiment, the reference points are distributed using a u×v uniform grid, covering the effective irradiation area aligned with the projection of the radiation source plane. The dimensions of the two planes are strictly matched, and the reference plane is set at Y = 100cm according to standard recommendations to simulate typical industrial irradiation conditions.

[0027] S13. Based on the radiation source plane and dose reference plane, and according to the standardized requirements for the spatial distribution of the radiation field in "JJG-591-1989 Gamma-ray Radiation Source (for Radiation Processing)," construct Co. 60 The three-dimensional parametric geometric model of the irradiation station, specifically as follows: Figure 2 As shown; S14, Dose field calculation; in the three-dimensional parametric geometric model, Co 60 The activity of the source rod can be considered approximately uniform, and its positional relationship with the irradiation point in a spatial rectangular coordinate system is as follows: Figure 3 As shown. The coordinates of the bottom endpoint of the source rod are (X... i , Y i Z i The coordinates of the top endpoint are (X... i , Y i Z i +L), based on this, calculate a single root Co 60 The radiation dose of the radiation source rod to the irradiation reference point O; in S14, the single Co rod is calculated. 60 The radiation dose from the radiation source rod to the irradiation reference point O is expressed as: In the formula, For the first Line 1 The radiation dose of the source rod to the irradiation reference point O, This is the irradiation constant, typically taken as 1.48 × 10⁻⁶. -4 Ci - ¹ cm²; Source rod activity, measured in Bq, represents the number of decaying nuclei per unit time; This represents the perpendicular distance between any irradiated point in the reference plane and the source rod. Indicates the length of the source rod. This represents the distance from the foot of the perpendicular to the lower vertex of the source rod; in, Specifically, it is expressed as follows: In the formula, , Let O be the coordinates of the irradiation reference point. , The coordinates of the source rod.

[0028] S15. Based on the radiation dose calculation results in S14, calculate all Co... 60 The total dose of the radiation source rod to a single reference point is calculated, and the dose non-uniformity of the reference plane is calculated. Among them, all Co 60 The total dose from a radiation source rod to a single reference point is a linear superposition of the dose contributions from each source rod, expressed as: In the formula, Total dose; Based on the total dose, the dose non-uniformity of the reference plane is calculated and expressed as: In the formula, Dose nonuniformity at the reference plane.

[0029] S16. In the newly constructed irradiation station scenario, the combinational degrees of freedom of the source rod arrangement reaches its theoretical maximum. Based on dose non-uniformity, Co... 60 The task of optimizing the arrangement of radiation source rods is formalized as a high-dimensional non-convex programming problem, and the search is performed to find the Co that minimizes the dose non-uniformity of the reference plane. 60 The arrangement and combination of radiation source rods is the optimization objective, which is expressed as: In the formula, This represents the minimum dose nonuniformity obtained under the optimal source rod arrangement scheme.

[0030] S2. Construct a pre-calculated dose coefficient matrix based on the radiation field fundamental model, and use the pre-calculated dose coefficient matrix as Co. 60 The dose assessment acceleration operator in the iterative process of radiation source rod arrangement optimization includes the following sub-steps: S21. During the algorithm initialization phase, according to Co... 60 The three-dimensional parametric geometric model of the irradiation station is used to construct a pre-calculated dose coefficient matrix, specifically by: [constructing the matrix based on the irradiation dose in S14]. The expression is separated into the geometric influence factor and the source rod activity A, and then the dose contribution matrix B is constructed. The total dose distribution D is then transformed into a matrix multiplication operation between the source rod activity A and the dose contribution matrix B, i.e.: ; The geometric influence factor is expressed as follows: In the formula, For dose contribution matrix The element in represents the slot number. OK The geometric influence of the source bar at column n on the reference point plane at row n and column m is determined by this matrix; a key characteristic of this matrix is ​​that... The calculation of the total dose distribution is entirely determined by the geometric parameters of the source rod (position, length, width) and the position of the reference point, and is independent of the source rod activity A. Therefore, during the optimization process, the activity A can be separated from the dose calculation formula, allowing the total dose distribution D to be calculated in real time using matrix multiplication.

[0031] S22, Dose Contribution Matrix B in Co 60 The radiation source rod arrangement optimization is calculated and stored once before the iteration, and can be repeatedly called in subsequent optimization iterations without recalculating the geometric influence factor, thus eliminating redundant calculation overhead.

[0032] Specifically, this embodiment pre-calculates and stores the dose contribution matrix B. Each arrangement adjustment only requires one matrix multiplication, eliminating the need to repeatedly calculate the geometric influence factor. This reduces the computational complexity of each iteration from O(U×V) to O(V), achieving a computational speedup of hundreds of times (experiments verify approximately 500 times). Compared to point-by-point calculation... The computational cost of matrix multiplication is significantly reduced, making it particularly suitable for optimization scenarios of large-scale irradiation stations.

[0033] In addition, matrix elements The method calculates and stores the dose contribution matrix B once before optimization, eliminating the need for repeated calculations in multiple optimization iterations and further reducing redundant computational overhead. This results in highly stable computation and broad applicability. However, this method assumes a static geometric configuration. If the source rod or reference point position changes dynamically, the dose contribution matrix B needs to be updated, potentially increasing computational costs. For large systems, storing the dose contribution matrix B may face memory limitations.

[0034] S3. Based on the dose assessment acceleration operator, an improved adaptive simulated annealing algorithm is used to process Co. 60 The radiation source rod arrangement scheme is subjected to global search and iterative optimization, and the final output is the Co value corresponding to the minimum dose inhomogeneity of the irradiation field. 60 The globally optimal arrangement of radiation source rods, refer to Figure 4 Specifically, it includes the following sub-steps: S31. Based on the source-slot coordinate parameters specified by the user, establish the initial spatial configuration of the irradiation system that matches the basic model of the radiation field; S32. Based on the dose reference plane defined in S12, set several reference points for dose assessment; S33, according to Co 60 Given the known activity distribution of the radiation source rods, generate an initial arrangement scheme for the source rods, and call the dose assessment acceleration operator to calculate the dose non-uniformity corresponding to the initial arrangement scheme; S34. Using the initial source rod arrangement as the initial solution for iteration, a new source rod arrangement is generated through a neighborhood search strategy, and the dose evaluation acceleration operator is called to calculate the dose non-uniformity of the new arrangement. S35. Determine whether to accept the new source bar arrangement scheme according to the Metropolis criterion. If the new solution is better, accept it directly; otherwise, accept the suboptimal solution with a preset probability. This ensures that the algorithm has sufficient global exploration capability and avoids getting trapped in local minima. The Metropolis criterion is as follows: In the formula, This represents the probability of accepting the new solution. and These represent the current solution and the optimized new solution, respectively. This represents the temperature parameter in the simulated annealing algorithm.

[0035] S36. Repeat S34~S35 for multiple rounds of iterative optimization until the algorithm converges, and output Co. 60 The globally optimal arrangement of radiation source rods; Specifically, combined with S2, by pre-storing the dose contribution of each source cell to each reference point, the dose evaluation during the optimization process can be transformed into matrix multiplication operations, thereby significantly accelerating the overall optimization process. The IASA algorithm employs an exponential annealing cooling strategy, maintaining strong global search capabilities in the initial stage and gradually converging to a local optimum in the later stage.

[0036] S4. Experimental Results and Analysis; To verify the effectiveness of the algorithm proposed in this invention, a Co used in actual production at a research institution was selected. 60The source rack serves as a reference. This source rack employs a three-layer structure (upper, middle, and lower layers), with each layer containing two sub-racks (left and right), forming a total of six sub-racks, named C, D, E, F, H, and K. Each sub-rack is equipped with 30 slots, for a total of 180 slots in the entire source rack. Slot naming follows an XY format, where X represents the sub-rack identifier (C, D, E, F, H, K), and Y defines the slot number (1 to 30) from left to right within each sub-rack. For example, the 13th slot from the left in sub-rack D is labeled D-13. In terms of geometry, the center-to-center distance between adjacent source rods within a sub-rack is 3.5 cm, the center-to-center distance between adjacent vertical source slots is 56 cm, and the center-to-center distance between adjacent horizontal source rods is 11 cm. Commonly used Co... 60 The source rod is 45.14 cm long and 1.11 cm in diameter. Figure 5 A schematic diagram of the source frame modeling; the verification method of this invention includes the following sub-steps: S41. To verify the comprehensive performance of the improved Adaptive Simulated Annealing (IASA) algorithm in newly created scenes, this invention is based on the Co described in S4. 60 Based on the physical parameters of the source rack, a three-dimensional dose field model was constructed. The algorithm input used the source rod activity data from the most recent source replacement operation at a research institute's irradiation facility, covering four batches of source rods from 2008, 2011, 2012, and 2025, with activity values ​​converted to September 16, 2025. The activity data is shown in Table 1.

[0037] Table 1 - Source Rod Numbers and Activity Step S42: Under the same hardware experimental environment (CPU: AMD Ryzen 7 7800X3D, GPU: NVIDIA RTX4070 Super, Memory: 32 GB DDR5-6400), the IASA algorithm proposed in this invention is compared with six other optimization algorithms, including the Divine Religions Algorithm (DRA), Hunger Games Search (HGS), Manta Ray Foraging Optimization (MRFO), Particle Swarm Optimization (PSO), Plant Growth Simulation Algorithm (PGSA), and Greedy Algorithm (GA). For the input activity information, each algorithm is run ten times, and its optimal dose inhomogeneity and running time are recorded for each run. Specific data are shown in Tables 2 and 3. For dose inhomogeneity, the target value for dose inhomogeneity must be less than 1.5.

[0038] Table 2 - Comparison of Algorithm Non-uniformity in New Scenarios Table 3 - Comparison of Algorithm Running Time in New Scenarios Step S43: To more intuitively compare the performance of each algorithm, based on the dose inhomogeneity and running time data from the above ten runs, the mean and standard deviation of each group of data were calculated. The calculation formula is shown below: In the formula, It is the mean of the data in each group. e1 is the standard deviation of each set of data, e2 is the value of the first run of each algorithm, en is the value of the nth run of each algorithm, and nnum is the number of runs. The calculation results are shown in Table 4. Table 4 - Comparison of Algorithm Parameters in New Scenarios Step S44, Reference Figure 6 To compare the performance of seven algorithms under the same activity data, a comparative graph was plotted based on dose inhomogeneity data from five independent runs: Experimental results based on actual production data show that the improved adaptive simulated annealing (IASA) algorithm proposed in this invention exhibits significant performance advantages in the newly constructed scenario, as shown in the table below: As shown in Tables 2, 3, and 4, the average dose nonuniformity of the IASA algorithm is significantly better than that of other comparative algorithms, fully verifying its excellent robustness and convergence accuracy. In terms of computational efficiency, the IASA algorithm achieves runtime performance comparable to the most advanced metaheuristic algorithms (such as DRA, HGS, and MRFO) through a dynamic neighborhood search strategy and matrix pre-computation technology. After ten independent experimental runs, the average runtime difference between the IASA algorithm and the fastest comparative algorithm is approximately 0.2 seconds. This performance advantage stems from three core design elements: (1) an adaptive annealing scheduling scheme based on dose gradient sensitivity, which can quickly identify high-quality solution domains at high temperatures; (2) hybrid neighborhood operations (including swapping, reversing, and block migration), which effectively overcome local optima obstacles; and (3) an incremental dose update strategy based on matrix masking, which significantly reduces the computational overhead of each iteration.

[0039] Experimental data strongly validates the engineering applicability of the IASA algorithm in multi-objective optimization of irradiation facilities, providing solid theoretical and practical support for intelligent decision-making on source rod layout in the nuclear industry.

[0040] S45: Comparison of Acceleration Matrix Operators To improve dose calculation efficiency, this invention proposes using a pre-computed dose coefficient matrix (DCM) as an acceleration operator. The DCM is constructed once before the optimization process, with each element representing the unit dose contribution from a specific source bar to a reference point. The construction formula is based on the spatial relationship between the geometric center of the source bar and the reference point, as well as the bar length, and is modeled using a simple arctan function. After constructing the matrix, the dose contribution during the optimization process is calculated through a single matrix multiplication of the activity vector and the DCM, thereby eliminating the need for real-time calculations and significantly reducing the time cost of each evaluation function call.

[0041] To verify the effectiveness of this acceleration method, experiments were conducted in a newly created scenario, comparing the performance of traditional methods and acceleration matrix methods under the same optimization workflow. The source rod activity data used are shown in Table 1. The Improved Adaptive Simulated (IASA) algorithm was executed five times for both DCM and DCM-free scenarios, and the mean and standard deviation of the optimized dose inhomogeneity and runtime were recorded. The comparison results are shown in Table 5. Table 5 - Comparison of acceleration performance of DCM in IASA algorithm Experimental results show that the improved adaptive simulated annealing (IASA) algorithm using the dose coefficient matrix (DCM) achieves comparable optimization results in dose nonuniformity to the IASA algorithm without DCM, while significantly outperforming traditional methods in terms of average running time. The acceleration factor approaches 500. This acceleration mechanism not only significantly improves optimization efficiency but also provides a feasible technical approach for real-time optimization in large-scale source rod deployment scenarios.

[0042] In summary, this invention provides a computational acceleration paradigm: by introducing a pre-computed dose coefficient matrix (DCM), the geometric factors in dose calculation are decoupled from the source rod activity. The geometric influence factor matrix is ​​calculated and stored once before optimization begins. During iteration, time-consuming real-time geometric calculations are transformed into efficient matrix multiplication operations (complexity reduced to O(V)), thereby solving the computational bottleneck of dose assessment and achieving a significant improvement in computational efficiency (experiments show an acceleration of hundreds of times), making rapid optimization of large-scale new scenarios possible.

[0043] We provide an efficient and robust global optimization algorithm: an improved adaptive simulated annealing (IASA) algorithm is designed. This algorithm enhances its global exploration and local refinement capabilities in a large solution space by integrating dynamic neighborhood search strategies (such as swapping, reversing, and block migration) and an adaptive annealing schedule, effectively avoiding premature convergence and ensuring stable searching for near-global optimal source rod arrangements with lower dose nonuniformity.

[0044] This invention integrates the aforementioned IASA algorithm with the DCM acceleration operator to construct an "algorithm-accelerator" collaborative optimization framework. This framework enables the optimization algorithm to perform more numerous and thorough searches while enjoying extremely fast single-iteration evaluation speed, thereby providing a more efficient evaluation of the newly created Co within an acceptable timeframe. 60 The irradiation facility outputs a high-quality, highly uniform source rod layout design, ultimately achieving a dual breakthrough in optimizing both efficiency and quality.

[0045] Although specific embodiments of the invention have been described in detail with reference to the accompanying drawings, this should not be construed as limiting the scope of protection of this patent. Various modifications and variations that can be made by a person skilled in the art without inventive effort within the scope described in the claims still fall within the scope of protection of this patent.

Claims

1. A method for optimizing the arrangement of Co-60 radiation source rods in a newly constructed scene based on DCM and IASA algorithms, characterized in that, Includes the following steps: S1, Construct Co 60 The basic model of the radiation field in the newly constructed scene of the radiation source rod; S2. Construct a pre-calculated dose coefficient matrix based on the aforementioned radiation field fundamental model, and use the pre-calculated dose coefficient matrix as Co. 60 A dose assessment acceleration operator in the iterative process of radiation source rod arrangement optimization; S3. Based on the aforementioned dose assessment acceleration operator, an improved adaptive simulated annealing algorithm is used to process Co. 60 The radiation source rod arrangement scheme is subjected to global search and iterative optimization, and the final output is the Co value corresponding to the minimum dose inhomogeneity of the irradiation field. 60 The globally optimal arrangement of radiation source rods.

2. The method for optimizing the arrangement of Co-60 radiation source rods in a newly constructed scene based on DCM and IASA algorithms according to claim 1, characterized in that, S1 includes the following sub-steps: S11. Define a radiation source plane in the XOZ coordinate system plane, and set a source slot in the radiation source plane to represent the placement position of the source rod. The coordinate set of the source slot is: ； In the formula, This represents the set of coordinates of the source slot within the radiation source plane; This represents the coordinates of the vertex on the source slot. Let x be the x-coordinate of the vertex on the source slot. Let be the vertical coordinate of the vertex on the source slot. , These represent the number of rows and columns of the source slot, respectively. and These represent the total number of rows and columns of the source slots, respectively. S12. Define a dose reference plane parallel to the radiation source plane. The dose reference plane is located at the vertical coordinate Y=100cm, and its coordinate set is as follows: ； In the formula, The set of coordinates representing the reference plane; Indicates the coordinates of the reference point. The x-coordinate of the reference point The vertical coordinate of the reference point, The row number of the reference point. The number of columns for the reference point. , These represent the total number of rows and columns of the reference point, respectively. S13. Based on the radiation source plane and dose reference plane, construct Co 60 A three-dimensional parametric geometric model of an irradiation station; S14. In a three-dimensional parametric geometric model, calculate a single root Co. 60 The radiation dose of the radiation source rod to the irradiation reference point O; S15. Based on the radiation dose calculation results in S14, calculate all Co... 60 The total dose of the radiation source rod to a single reference point is calculated, and the dose non-uniformity of the reference plane is calculated. S16. In the newly constructed irradiation station scenario, based on dose non-uniformity, Co... 60 The task of optimizing the arrangement of radiation source rods is formalized as a high-dimensional non-convex programming problem.

3. The method for optimizing the arrangement of newly constructed Co-60 radiation source rods based on DCM and IASA algorithms according to claim 2, characterized in that, In S14, the calculation of a single Co root 60 The radiation dose from the radiation source rod to the irradiation reference point O is expressed as: ； In the formula, For the first Line 1 The radiation dose of the source rod to the irradiation reference point O, The irradiation constant, For source rod activity, This represents the perpendicular distance between any irradiated point in the reference plane and the source rod. Indicates the length of the source rod. This represents the distance from the foot of the perpendicular to the lower vertex of the source rod.

4. The method for optimizing the arrangement of newly constructed Co-60 radiation source rods based on DCM and IASA algorithms according to claim 3, characterized in that, In S15, all Co are calculated. 60 The total dose from the radiation source rod to a single reference point is expressed as: ； In the formula, Total dose; Based on the total dose, the dose non-uniformity of the reference plane is calculated, and it is expressed as: ； In the formula, Dose nonuniformity at the reference plane.

5. The method for optimizing the arrangement of newly constructed Co-60 radiation source rods based on DCM and IASA algorithms according to claim 4, characterized in that, In S16, Co 60 The task of optimizing the arrangement of radiation source rods is formalized as a high-dimensional non-convex programming problem, and the search is performed to find the Co that minimizes the dose non-uniformity of the reference plane. 60 The arrangement and combination of radiation source rods is the optimization objective, which is expressed as: ； In the formula, This represents the minimum dose nonuniformity obtained under the optimal source rod arrangement scheme.

6. The method for optimizing the arrangement of newly constructed Co-60 radiation source rods based on DCM and IASA algorithms according to claim 3, characterized in that, S2 includes the following sub-steps: S21, The pre-calculated dose coefficient matrix is ​​constructed by: taking the irradiation dose in S14... The expression is separated into the geometric influence factor and the source rod activity A, and then the dose contribution matrix B is constructed. The total dose distribution D is then transformed into a matrix multiplication operation between the source rod activity A and the dose contribution matrix B, i.e.: ; S22, Dose Contribution Matrix B in Co 60 The radiation source rod arrangement optimization is calculated and stored once before the iteration, and can be repeatedly called in subsequent optimization iterations without recalculating the geometric influence factor, thus eliminating redundant calculation overhead.

7. The method for optimizing the arrangement of Co-60 radiation source rods in a newly constructed scene based on DCM and IASA algorithms according to claim 6, characterized in that, In S21, the geometric influence factor is expressed as follows: ； In the formula, For dose contribution matrix The element in represents the slot number. OK The geometric influence factor of the source bar at column n on the reference point plane at row m.

8. The method for optimizing the arrangement of newly constructed Co-60 radiation source rods based on DCM and IASA algorithms according to claim 2, characterized in that, S3 includes the following sub-steps: S31. Based on the source-slot coordinate parameters specified by the user, establish the initial spatial configuration of the irradiation system that matches the basic model of the radiation field; S32. Based on the dose reference plane defined in S12, set several reference points for dose assessment; S33, according to Co 60 Given the known activity distribution of the radiation source rods, an initial arrangement scheme of the source rods is generated, and the dose assessment acceleration operator is called to calculate the dose non-uniformity corresponding to the initial arrangement scheme. S34. Using the initial source rod arrangement scheme as the initial solution for iteration, a new source rod arrangement scheme is generated through a neighborhood search strategy, and the dose evaluation acceleration operator is called to calculate the dose non-uniformity of the new arrangement scheme. S35. Determine whether to accept the new source bar arrangement scheme according to the Metropolis criterion. If the new solution is better, accept it directly; otherwise, accept the suboptimal solution with a preset probability. S36. Repeat S34~S35 for multiple rounds of iterative optimization until the algorithm converges, and output Co. 60 The globally optimal arrangement of radiation source rods.

9. The method for optimizing the arrangement of newly constructed Co-60 radiation source rods based on DCM and IASA algorithms according to claim 8, characterized in that, In S35, the Metropolis criterion is: ； In the formula, This represents the probability of accepting the new solution. and These represent the current solution and the optimized new solution, respectively. This represents the temperature parameter in the simulated annealing algorithm.

Images