Microwave hyperthermia phased array excitation optimization method based on adaptive multi-target particle swarm algorithm
By optimizing the microwave hyperthermia phased array excitation through an adaptive multi-objective particle swarm algorithm, the problems of focus deviation and insufficient focusing range in microwave hyperthermia technology are solved, precise focusing and reduction of redundant hot spots are achieved in a shorter time, and the safety and resolution of treatment are improved.
Patent Information
- Application Number
- CN202510715754.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-30
- Publication Date
- 2025-09-30
AI Technical Summary
Existing microwave hyperthermia technology has problems such as focus deviation from the target, excessive focusing range and insufficient resolution when focusing on tumors. In addition, existing methods fail to effectively combine intelligent optimization algorithms with reasonable objective functions, resulting in long calculation time and insufficient treatment safety.
An adaptive multi-objective particle swarm optimization algorithm is used to optimize the excitation of microwave hyperthermia phased array. By setting multiple optimization objectives and specific objective functions, combined with the particle swarm optimization algorithm, the excitation of the array antenna units is optimized to achieve precise focusing and reduce unnecessary hotspots.
Precise focusing is achieved in a shorter computing time, unnecessary hot spots are reduced, the safety of treatment and focusing resolution are improved, and the damage rate of healthy tissue is reduced.
Smart Images
Figure CN120727196A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of phased arrays, and in particular relates to a microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm algorithm. Background Art
[0002] With advances in cancer treatment technology, hyperthermia has become a promising alternative to radiotherapy and chemotherapy. Compared with traditional treatments, hyperthermia can significantly improve local tumor control rates and minimize adverse reactions to treatment. Therefore, it has been widely used in clinical trials. The purpose of hyperthermia is to concentrate energy on the tumor to inactivate it. Compared with other hyperthermia treatments, microwave hyperthermia technology uses phased arrays to generate focused microwaves. Due to its advantages such as fast ablation speed, accurate positioning, ease of use, low risk of postoperative complications, and low medical costs, it has attracted widespread attention in the academic community. The main computational challenge of microwave hyperthermia is how to accurately concentrate microwave energy on tumor tissue. However, during the process of optimizing treatment, microwave attenuation and scattering can cause the focus to deviate from the target.
[0003] To address the above issues, an important approach is to optimize the excitation of the array antenna units. Although existing research methods have achieved focusing on the tumor, the focusing range is much larger than the tumor area, and the focusing resolution is significantly insufficient. Some current methods choose to increase the number of elements in the array antenna to improve resolution, but the cost is that the focus offset and redundant hot spots are more likely to occur. In addition, existing implementation methods do not consider making corresponding improvements to the intelligent optimization algorithm and combining it with a reasonable objective function to optimize the array feed, thereby achieving precise focusing in a shorter computing time and better reducing redundant hot spots, thereby improving the safety of treatment. Summary of the Invention
[0004] Purpose of the invention: To address the problems that existing methods are insufficient in efficiency and adaptability, and that conventional hotspot target entropy, average power absorption ratio, and specific absorption rate lack accuracy when used as objective functions to describe hotspots, the present invention proposes a microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm algorithm, which achieves precise focusing and reduces redundant hotspots in a shorter computing time, resulting in higher treatment safety.
[0005] Technical solution: A microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm algorithm, comprising the following steps:
[0006] Step 1: Treat each antenna unit in the phased array as an individual, establish a corresponding population, set an optimization target for the population, and initialize the position and velocity of each individual. The individual position includes the phase and amplitude of the antenna unit excitation. The optimization target includes: hotspot offset = 0mm, number of hotspots = 1 and treatment quantitative index; the hotspot offset is the distance between the strongest hotspot position and the target position, the strongest hotspot is the area where the maximum specific absorption rate is located, and the strongest hotspot position is its corresponding position. The hotspot represents the area where the specific absorption rate of human tissue is greater than the maximum value. The area formed;
[0007] Step 2: Simulate the electromagnetic energy field distribution in the human body according to the position of each individual, and calculate the specific absorption rate based on the electromagnetic energy field distribution in the human body;
[0008] Step 3: Calculate the objective function based on the optimization target and specific absorption rate;
[0009] Step 4: According to the objective function value, search for non-inferior solutions and update the optimal value of the population;
[0010] Step 5: Adaptively adjust the speed of the individual according to the difference between the optimal solution of the population and the optimization target, and update the individual position according to the adjusted individual speed;
[0011] Step 6: Determine whether the optimal solution of the population meets the optimization goal or the number of iterations reaches the upper limit. If so, output the optimal solution, otherwise execute step 2;
[0012] Step 7: Adjust the phased array excitation based on the optimal solution.
[0013] Furthermore, the treatment quantitative indicators include: power absorption peak ratio <0.5, hotspot target entropy, average power absorption ratio, tumor coverage >98%, and healthy tissue damage rate <5%;
[0014] The tumor coverage is the percentage of the tumor volume in the tissue whose specific absorption rate is greater than the maximum value n%.
[0015] Furthermore, based on the distribution of electromagnetic energy fields in the human body, the specific absorption rate is calculated according to the following formula:
[0016]
[0017] Where ρ(r) is the density of human tissue, σ(r) is the conductivity of human tissue, is the total electric field inside the human tissue, and r is the distance vector from the observation point to the array center.
[0018] Furthermore, the objective function is expressed as:
[0019] min(error,pPA,1 / aPA,HTQ,1 / (TC 70 +1),DHTR)
[0020] Where min(·) represents minimization, error is the hotspot offset, n is the number of hotspots, pPA is the peak power absorption ratio, aPA is the average power absorption ratio, HTQ is the hotspot target entropy, TC is the target entropy, and 70 is the tumor coverage rate, and DHTR represents the healthy tissue damage rate.
[0021] Furthermore, the power absorption peak ratio is expressed as:
[0022]
[0023] Where, for each SAR exceeding the maximum A continuous area, the highest specific absorption rate in the area is defined as the local peak, SAR peakmax is the maximum value among the local peaks, SAR peak (i) For reasons other than SAR peakmax An array of local peaks other than SAR, i is the SAR peak The number in the array.
[0024] Furthermore, the average power absorption ratio is expressed as:
[0025]
[0026] Among them, P tumor and P healthy Denote the power absorbed by tumor and healthy tissue respectively, V tumor and V healthy are the volumes of tumor and healthy tissue, respectively.
[0027] Furthermore, the hotspot target entropy is expressed as:
[0028]
[0029] Where V1 is the volume of the top 1% of the specific absorption rate values in healthy tissue, arranged from strong to weak. is the average specific absorption rate of V1, is the average specific absorption rate of the tumor.
[0030] Furthermore, the tumor coverage is expressed as:
[0031]
[0032] Where V tumor is the volume of tumor and healthy tissue;
[0033] The healthy tissue damage rate is expressed as:
[0034]
[0035] Where V healthy is the volume of healthy tissue, when the power absorption of healthy tissue exceeds the maximum value When the damage is caused, V damage Represents the volume of healthy tissue that is damaged.
[0036] Furthermore, in step 5, the speed of the individual is adaptively adjusted according to the difference between the optimal solution of the population and the optimization target. The specific operations include:
[0037] According to the difference between the optimal solution of the population and the optimization target, the speed inertia weight is adaptively updated according to the following formula:
[0038]
[0039] Where f is the difference mapping between the optimal solution of the population and the optimization target, and rand is a random number uniformly distributed in the range [0,1].
[0040] According to the difference between the optimal solution of the population and the optimization target, the self-learning factor is adaptively updated according to the following formula:
[0041]
[0042] According to the difference between the optimal solution of the population and the optimization target, the social learning factor is adaptively updated according to the following formula:
[0043]
[0044] The updated speed is expressed as:
[0045]
[0046] in, is the j-dimensional velocity of the i-th individual in the t-th iteration population, w is the velocity inertia weight, c1 is the self-learning factor, c2 is the social learning factor, is a local optimal solution, is the global optimal solution, and are two random numbers uniformly distributed in the range [0,1]. is the j-th dimension position of the i-th individual in the t-th iteration population.
[0047] Furthermore, in step 5, the individual position is updated according to the adjusted individual speed. The specific operations include:
[0048] Update the individual's position as follows:
[0049]
[0050] Where, is the j-th dimension position of the i-th individual in the t+1-th iteration population, is the j-th dimension velocity of the i-th individual in the t+1-th iteration population.
[0051] Beneficial effects: Compared with the existing technology, the present invention considers multiple optimization objectives and simultaneously evaluates multiple treatment quantitative indicators in solving the focusing problem of the thermal therapy phased array. By combining the particle swarm optimization algorithm and a specific objective function to optimize and improve the electromagnetic field energy distribution, it can achieve precise focusing in a shorter computing time, reduce redundant hot spots and reduce the damage rate of healthy tissue. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] Figure 1 is a flow chart of the steps of the present invention;
[0053] Figure 2 A model of the present invention;
[0054] Figure 3 Optimization process diagram of the present invention;
[0055] Figure 4 Schematic diagram for comparing the results of the present invention, wherein: Figure 4 (a) is the result obtained by using the standard particle swarm optimization algorithm. Figure 4 (b) is the result obtained by using the linear decreasing weight particle swarm optimization algorithm. Figure 4 (c) in the figure is the result obtained by using genetic algorithm. Figure 4 (d) in the figure is the result obtained by using the differential evolution algorithm. Figure 4 (e) is the result obtained by the method of the present invention. DETAILED DESCRIPTION
[0056] The technical solution of this embodiment will now be further described with reference to the accompanying drawings and embodiments.
[0057] Example 1:
[0058] like Figure 1 As shown, this embodiment proposes a microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm algorithm, which includes the following steps:
[0059] Step 1: Establish a corresponding population based on the number of phased array antenna units, set optimization targets for the population, and initialize individuals in the population. In this embodiment, the population established contains 16 individuals, each of which carries 32-dimensional feasible solution information, and limits the maximum number of iterations to 50. For a tumor located at (0, 35, 35) mm, the optimization targets include hotspot offset = 0 mm, number of hotspots = 1, and treatment quantification indicators. Treatment quantification indicators include power absorption peak ratio < 0.5, hotspot target entropy < 0.5, average power absorption ratio > 13, tumor coverage > 98%, and healthy tissue damage rate < 5%. Randomly generate initialization positions and velocities for the individuals. The positions include feasible solutions for the phase and amplitude of the unit excitation, with the phase limited to [-180°, 180°] and the amplitude limited to [0, 1].
[0060] Step 2: Input the individual position correspondence into the phase and amplitude of the array feed for simulation, and calculate the specific absorption rate (SAR) based on the electromagnetic energy field distribution. The calculation formula is:
[0061]
[0062] Where ρ(r) represents the density of human tissue, σ(r0) represents the conductivity of human tissue, represents the total electric field inside human tissue, and r is the distance vector from the observation point to the center of the array.
[0063] Calculate the objective function, which is expressed as:
[0064] min(error,pPA,1 / aPA,HTQ,1 / (TC 70 +1),DHTR)
[0065] Among them, min(g) in the objective function is the minimization target, error is the hotspot offset, n is the number of hotspots, pPA is the peak power absorption ratio, aPA is the average power absorption ratio, HTQ is the hotspot target entropy, TC 70 is the tumor coverage rate, DHTR represents the healthy tissue damage rate. The hotspot offset is the distance between the strongest hotspot position and the target position. The strongest hotspot is the area where the specific absorption rate is the maximum, and the position is the position corresponding to the maximum. The number of hotspots is the number of tissues with a specific absorption rate greater than the maximum. The number of regions.
[0066] The calculation formula for the power absorption peak ratio pPA is:
[0067]
[0068] Where, for each specific absorption rate exceeding the maximum value A continuous area, the highest specific absorption rate in the area is defined as the local peak, SAR peak is the local peak value obtained based on the specific absorption rate, SAR peakmax is the maximum value among the local peaks, SAR peak (i) For SAR peak SAR peakmax An array of values other than SAR, i is SAR peak The number in the array.
[0069] Among them, the calculation formula of average power absorption ratio aPA is:
[0070]
[0071] Where P tumor and P healthy Denote the power absorbed by tumor and healthy tissue respectively, V tumor and V healthy are the volumes of tumor and healthy tissue, respectively.
[0072] Among them, the calculation formula of hot spot target entropy HTQ is:
[0073]
[0074] Where V1 is the volume of the top 1% of the specific absorption rate values in healthy tissue, arranged from strong to weak. is the average SAR value in V1, is the average SAR value in the tumor.
[0075] Among them, tumor coverage TC n is the percentage of tumor volume in the tissue with a specific absorption rate greater than n% of the maximum value, and the calculation formula is:
[0076]
[0077] The calculation formula for healthy tissue damage rate DHTR is:
[0078]
[0079] Where V healthy is the volume of healthy tissue, when the power absorption of healthy tissue exceeds the maximum value When the damage is caused, so V damage Represents the volume of healthy tissue that is damaged.
[0080] To search for non-inferior solutions and update the optimal value, first determine whether the individual in this iteration is a non-inferior solution to the historical results. If the individual is better than the historical results in at least one objective and does not degrade in other objectives, the solution is a non-inferior solution and is updated; otherwise, the historical results are retained. Then, the optimal solution of the population is found and determined to be a non-inferior solution to the historical results. If it is a non-inferior solution, it is updated; otherwise, the historical results are retained.
[0081] Step 3: Adaptively adjust the update speed of the individual according to the difference between the optimal solution of the population and the target, and update the position of the individual in the population. The update speed is expressed as:
[0082]
[0083] in, is the j-dimensional velocity of the i-th individual in the t-th iteration population, w is the velocity inertia weight, c1 is the self-learning factor, c2 is the social learning factor, is a local optimal solution, is the global optimal solution, and are two random numbers uniformly distributed in the range [0,1]. is the j-th dimension position of the i-th individual in the t-th iteration population.
[0084] In this embodiment, the adaptive adjustment targets are the speed inertia weight, self-learning factor, and social learning factor in the speed expression. The adaptive adjustment function adopts a nonlinear, fluctuating value form, which is manifested by the addition of random numbers to the function. The adaptive adjustment strategy is: as the difference between the optimal solution of the population and the target decreases, the volatility of the updated speed inertia weight and self-learning factor decreases, while the volatility of the social learning factor increases. To achieve this, the difference between the optimal solution of the population and the target is calculated, and then the speed inertia weight, self-learning factor, and social learning factor are updated.
[0085] Among them, the adaptive function of velocity inertia weight is:
[0086]
[0087] Among them, the adaptive function of the self-learning factor is:
[0088]
[0089] Among them, the adaptive function of the social learning factor is:
[0090]
[0091] Among them, f (f>0) is the difference mapping between the optimal solution of the population and the target, and rand is a random number uniformly distributed in the range [0,1].
[0092] Update the particle position as follows:
[0093]
[0094] Where, is the j-th dimension position of the i-th individual in the t+1-th iteration population, is the j-th dimension velocity of the i-th individual in the t+1-th iteration population.
[0095] For individuals in an outlier state, that is, particles that have failed to update their individual optimal solutions for three consecutive iterations, it is necessary to redistribute the speed based on the difference between the individual's current objective function value and the target value. The larger the difference, the faster the speed. In this embodiment, the speed redistribution function is:
[0096]
[0097] Among them, pPA best , pPA target is the optimal value and target value of the current iteration power absorption peak ratio, error best 、error target The optimal and target values for the focus offset of the current iteration.
[0098] Step 4: Iterate and update the population until the optimal solution of the population meets the optimization goal or the number of iterations reaches the upper limit. If it meets the goal, output the optimization result. If not, return to step 2 for iteration.
[0099] Example 2:
[0100] The microwave hyperthermia phased array excitation optimization method based on the adaptive multi-objective particle swarm algorithm proposed in Example 1 is now applied to the optimization treatment of a heterogeneous simplified breast model.
[0101] like Figure 2As shown, the heterogeneous simplified breast model used in this embodiment consists of skin 1, fat 2, breast fibrosis 3, and a tumor 4. Its dielectric and thermal properties at 2.45 GHz are shown in Table 1. The heterogeneous simplified breast model is placed in a circular phased array consisting of 16 evenly arranged antenna elements 5. The circular phased array has a radius of 100 mm and a height of 200 mm. The phased array element measures 31.25 mm × 21.4 mm and uses a Rogers RT5880 with a dielectric constant of 2.2, a loss tangent of 0.009, and a thickness of 1.575 mm. The operating frequency range is 2-3 GHz, and the performance achieved at 2.45 GHz is -36 dB, demonstrating excellent performance. The phased array and heterogeneous simplified breast model are immersed in a coupling fluid (oil-water mixture 6) with a relative dielectric constant of 22.9 and a conductivity of 0.07. The tumor is located at (0, 35, 35) mm relative to the center of the array.
[0102] Table 1 Dielectric and thermal properties of the inhomogeneous simplified breast model at 2.45 GHz
[0103]
[0104]
[0105] The optimization process is completed by Matlab R2021a and HFSS. Figure 3 As shown in the figure, HFSS simulation can be used to obtain the electromagnetic energy field distribution and calculate the specific absorption rate distribution data. Subsequently, data processing and analysis in Matlab can be performed to obtain characteristic parameters. Based on these parameters, the phased array excitation is iteratively optimized and adjusted until the predetermined objectives are met, ultimately achieving appropriate element excitation.
[0106] Under the same hardware and software conditions, programming simulation was performed using the method of the present invention, the standard particle swarm optimization algorithm, the linear decreasing weight particle swarm optimization algorithm, the genetic algorithm, and the differential evolution algorithm. The optimized treatment quantitative indicators were recorded as shown in Table 2.
[0107] Table 2 Quantitative indicators of treatment after optimization
[0108]
[0109] The calculation methods for each treatment quantitative index are as follows:
[0110] The focal shift error is the distance between the center of the strongest hotspot and the center of the tumor.
[0111] The number of hot spots n is the number of valid focal points inside the breast model.
[0112] The expression of power absorption peak ratio pPA is:
[0113]
[0114] The calculation formula of hot spot target entropy HTQ is:
[0115]
[0116] Where V1 is the volume of the top 1% of the specific absorption rate values in healthy tissue, arranged from strong to weak. is the average SAR value in V1, is the average SAR value in the tumor.
[0117] The calculation formula of average power absorption ratio aPA is:
[0118]
[0119] Among them, P tumor and P healthy Denote the power absorbed by tumor and healthy tissue respectively, V tumor and V healthy are the volumes of tumor and healthy tissue, respectively.
[0120] Tumor coverage TC n The calculation formula is:
[0121]
[0122] TC n The SAR value in the tissue is greater than the maximum SAR of the tumor volume percentage n%. By observing the coverage of different indicators n, the energy distribution inside the tumor can be observed. This paper also evaluates TC 50 TC 70 and TC 90 effect.
[0123] The calculation formula of healthy tissue damage rate DHTR is:
[0124]
[0125] Where V healthy is the volume of healthy tissue, when the power absorption of healthy tissue exceeds the maximum value When the damage is caused, so V damage Represents the volume of healthy tissue that is damaged.
[0126] Compare pPA, HTQ, aPA, and TC in Table 2 90 TC 70 TC 50and DHTR indicators, the algorithm of the present invention is far superior to other algorithms. Compared with other optimization methods, the algorithm of the present invention found the ideal result within 50 iterations, achieved precise focusing and successfully reduced excess hotspots. The corresponding pPA dropped to 0.4431, indicating that the power energy is highly concentrated inside the tumor. In contrast, the standard particle swarm optimization algorithm cannot achieve the ideal focusing effect within a limited number of iterations. The focusing effect of the linear decreasing weight particle swarm optimization algorithm and the genetic algorithm is better than the standard particle swarm optimization algorithm, but there is still focusing offset and it is unable to reduce excess hotspots. The differential evolution algorithm achieves precise focusing, but cannot reduce excess hotspots.
[0127] See Figure 4 Comparison of optimization results shows that the algorithm of the present invention achieves precise focusing and keeps the number of hotspots to a single one. No other hotspots appear outside the focal plane. However, the standard particle swarm optimization algorithm exhibits significant focus shift. The linear decrement particle swarm optimization algorithm and the differential evolution algorithm exhibit a significant number of hotspots. The genetic algorithm exhibits focus shift and exhibits a significant number of hotspots outside the observation plane.
[0128] Compared to existing approaches to thermal therapy phased array focusing, the algorithm-optimized array feed of the present invention achieves precise focusing and reduces unwanted hotspots. The present invention's microwave thermal therapy phased array excitation optimization method, based on an adaptive multi-objective particle swarm algorithm, is accurate and effective. The optimized electromagnetic field is well distributed, and the optimization process is less prone to premature convergence. This method enables precise focusing, reduces unwanted hotspots, and lowers the rate of healthy tissue damage in a shorter computation time.
Claims
1. A microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm optimization algorithm, characterized by: The following steps are involved: Step 1: Treat each antenna unit in the phased array as an individual, establish a corresponding population, set an optimization target for the population, and initialize the position and velocity of each individual. The individual position includes the phase and amplitude of the antenna unit excitation. The optimization target includes: hotspot offset = 0mm, number of hotspots = 1 and treatment quantitative index; the hotspot offset is the distance between the strongest hotspot position and the target position, the strongest hotspot is the area where the maximum specific absorption rate is located, and the strongest hotspot position is its corresponding position. The hotspot represents the area where the specific absorption rate of human tissue is greater than 1 / the maximum value. the area formed; Step 2: Simulate the electromagnetic energy field distribution in the human body according to the position of each individual, and calculate the specific absorption rate based on the electromagnetic energy field distribution in the human body; Step 3: Calculate the objective function based on the optimization target and specific absorption rate; Step 4: According to the objective function value, search for non-inferior solutions and update the optimal value of the population; Step 5: Adaptively adjust the speed of the individual according to the difference between the optimal solution of the population and the optimization target, and update the individual position according to the adjusted individual speed; Step 6: Determine whether the optimal solution of the population meets the optimization goal or the number of iterations reaches the upper limit. If so, output the optimal solution, otherwise execute step 2; Step 7: Adjust the phased array excitation based on the optimal solution.
2. The microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm optimization algorithm according to claim 1, characterized in that: The treatment quantitative indicators include: power absorption peak ratio <0.5, hotspot target entropy, average power absorption ratio, tumor coverage >98%, and healthy tissue damage rate <5%; The tumor coverage is the percentage of the tumor volume in the tissue whose specific absorption rate is greater than the maximum value n%.
3. The microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm optimization algorithm according to claim 2, characterized in that: Based on the distribution of electromagnetic energy field in the human body, the specific absorption rate is calculated according to the following formula: Where ρ(r) is the density of human tissue, σ(r) is the conductivity of human tissue, is the total electric field inside the human tissue, and r is the distance vector from the observation point to the array center.
4. The microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm optimization algorithm according to claim 3, characterized in that: The objective function is expressed as: min(error,pPA,1 / aPA,HTQ,1 / (TC 70 +1),DHTR) Where min(·) represents minimization, error is the hotspot offset, n is the number of hotspots, pPA is the peak power absorption ratio, aPA is the average power absorption ratio, HTQ is the hotspot target entropy, TC is the target entropy, and 70 is the tumor coverage rate, and DHTR represents the healthy tissue damage rate.
5. The microwave hyperthermia phased array excitation optimization method based on adaptive multi-objective particle swarm optimization according to claim 4, characterized in that: The power absorption peak ratio is expressed as: Where, for each SAR exceeding the maximum A continuous area, the highest specific absorption rate in the area is defined as the local peak, SAR peakmax is the maximum value among the local peaks, SAR peak (i0 is the reason for excluding SAR peakmax An array of local peaks other than SAR, i is the SAR peak The number in the array.
6. The microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm optimization algorithm according to claim 4, characterized in that: The average power absorption ratio is expressed as: Among them, P tumor and P healthy Denote the power absorbed by tumor and healthy tissue respectively, V tumor and V healthy are the volumes of tumor and healthy tissue, respectively.
7. The microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm optimization algorithm according to claim 4, characterized in that: The hotspot target entropy is expressed as: Where V1 is the volume of the top 1% of the specific absorption rate values in healthy tissue, arranged from strong to weak. is the average specific absorption rate of V1, is the average specific absorption rate of the tumor.
8. The microwave hyperthermia phased array excitation optimization method based on an adaptive multi-objective particle swarm optimization algorithm according to claim 4, characterized in that: The tumor coverage is expressed as: Where V tumor is the volume of tumor and healthy tissue; The healthy tissue damage rate is expressed as: Where V healthy is the volume of healthy tissue, when the power absorption of healthy tissue exceeds the maximum value When the damage is caused, V damage Represents the volume of healthy tissue that is damaged.
9. The microwave hyperthermia phased array excitation optimization method based on adaptive multi-objective particle swarm optimization according to claim 1, characterized in that: In step 5, the speed of the individual is adaptively adjusted based on the difference between the optimal solution of the population and the optimization target. The specific operations include: According to the difference between the optimal solution of the population and the optimization target, the speed inertia weight is adaptively updated according to the following formula: Where f is the difference mapping between the optimal solution of the population and the optimization target, and rand is a random number uniformly distributed in the range [0,1]. According to the difference between the optimal solution of the population and the optimization target, the self-learning factor is adaptively updated according to the following formula: According to the difference between the optimal solution of the population and the optimization target, the social learning factor is adaptively updated according to the following formula: The updated speed is expressed as: in, is the j-dimensional velocity of the i-th individual in the t-th iteration population, w is the velocity inertia weight, c1 is the self-learning factor, c2 is the social learning factor, is a local optimal solution, is the global optimal solution, and are two random numbers uniformly distributed in the range [0,1]. is the j-th dimension position of the i-th individual in the t-th iteration population.
10. The microwave hyperthermia phased array excitation optimization method based on adaptive multi-objective particle swarm optimization according to claim 1, characterized in that: In step 5, the individual position is updated according to the adjusted individual speed. The specific operations include: Update the individual's position as follows: Where, is the j-th dimension position of the i-th individual in the t+1-th iteration population, is the j-th dimension velocity of the i-th individual in the t+1-th iteration population.