Medical instrument parameter intelligent optimization setting method based on evolutionary algorithm, storage medium and computer equipment

Abstract

The invention belongs to the technical field of intelligent medical treatment, and particularly relates to a medical instrument parameter intelligent optimization setting method based on an evolutionary algorithm, a storage medium and computer equipment, and the method comprises the steps: analyzing a medical instrument parameter optimization problem, determining a target function and a constraint condition, and building an optimization problem mathematical model; initializing the population of the parameters of the medical instrument to be optimized to form a parent population P of NP individuals; based on a crossover operator, generating a filial generation population oP sharing NP individuals; on the basis of a mutation operator, mutation operation is carried out on NP individuals of the offspring population, and a mutated offspring population is obtained; combining the parent population P with the varied offspring population to form a new population Q; calculating the fitness of each individual in the new population Q based on a self-balancing fitness function, and screening NP individuals with the minimum fitness to form a new parent population; and obtaining a group of optimal medical instrument parameters until a termination condition is reached. According to the invention, the safety and efficiency of digestive tract polyp resection are improved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent medical technology, and in particular relates to a medical instrument parameter intelligent optimization setting method based on an evolutionary algorithm, a storage medium and a computer device. Background Art

[0002] Gastrointestinal polypectomy is one of the most common minimally invasive surgeries in gastroenterology, and is widely used to treat benign and early malignant lesions of the digestive tract. Clinical instruments such as laser cutting and high-frequency electrosurgical units have gradually become important means of clinical application due to their advantages such as simple operation, less bleeding, and fast postoperative recovery.

[0003] However, in current clinical operations, the parameter settings of clinical instruments are subjective, overly dependent on the doctor's experience, and have poor parameter applicability, which leads to the risk of damaging surrounding healthy tissues and postoperative complications. With the development of science and technology, the medical industry is transforming and upgrading to smart medicine. The intelligent optimization setting of clinical medical instrument parameters is the basis for future intelligent gastrointestinal polyp resection surgery, and has important clinical value for the safety and quality of gastrointestinal polyp resection surgery.

[0004] With the development of intelligent optimization algorithms, the problem of intelligent optimization setting of parameters has been studied at home and abroad. Among them, the parameter optimization setting method based on penalty function regards the degree of constraint violation as part of the fitness function, which is simple and effective; however, the penalty function has many parameters and needs to be constantly adjusted to adapt to different optimization problems. When optimizing clinical instrument parameters, it is difficult to apply to different patient conditions. The method based on separation of objectives and constraints treats objectives and constraints separately, and the constraint processing efficiency is high; however, this type of method will over-relax some constraints, making it difficult to adaptively balance the feasibility and convergence of the optimal solution, resulting in the difficulty in ensuring the accuracy and safety of the surgery. The method based on multi-objective function regards the constraints as one or more objective functions, and the solution efficiency is high; however, this type of method often cannot guarantee the feasibility of the solution and cannot be widely used in medicine.

[0005] Therefore, how to combine medical experience and use intelligent optimization algorithms to quickly and accurately set feasible clinical medical instrument parameters in different situations in order to ensure cutting accuracy, reduce damage to surrounding healthy tissues, and reduce the risk of postoperative complications is an issue that urgently needs to be addressed in the field of gastroenterology. Summary of the invention

[0006] In view of this, the present invention provides a method, storage medium and computer equipment for intelligent optimization setting of medical instrument parameters based on evolutionary algorithm, which improves the safety, efficiency and intelligence of gastrointestinal polypectomy and solves the problems of poor applicability of algorithm parameters and difficulty in searching for feasible solutions in the prior art.

[0007] The intelligent optimization setting method in the embodiment of the present invention comprises the following steps:

[0008] S1. Analyze the medical instrument parameter optimization problem, determine the objective function and constraints, and establish a mathematical model for the optimization problem;

[0009] S2, initializing the population of medical instrument parameters to be optimized to form a parent population P with a total of NP individuals;

[0010] S3. Based on the crossover operator, a total of NP individuals in the offspring population oP is generated; based on the mutation operator, the NP individuals in the offspring population are mutated to obtain the mutated offspring population; the parent population P is merged with the mutated offspring population to form a new population Q;

[0011] S4. Based on the self-balancing fitness function, the fitness of each individual in the new population Q is calculated, and NP individuals with the smallest fitness are selected to form a new parent population;

[0012] S5. Repeat steps S3 and S4 until a termination condition is reached to obtain a set of optimal medical instrument parameters.

[0013] The storage medium in the embodiment of the present invention stores computer instructions thereon, and when the computer instructions are executed by the processor, each step of the intelligent optimization setting method is implemented.

[0014] The computer device in an embodiment of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor runs the computer program, the steps of the intelligent optimization setting method are executed.

[0015] It can be seen from the above technical solution that compared with the prior art, the present invention has the following beneficial effects:

[0016] The present invention uses a self-balancing fitness function to adaptively balance feasibility and diversity in an infeasible stage; in a semi-feasible stage, the balance between feasibility and diversity is shifted to a balance between convergence and diversity; in a feasible stage, the convergence and diversity are adaptively balanced. Therefore, the present invention can adapt to different optimization problems and has strong applicability.

[0017] In addition, the optimization algorithm of the present invention not only has fewer parameters, but also can ensure the feasibility and accuracy of the optimal solution, thereby ensuring the quality and safety of the operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of a flow chart of a method for intelligently optimizing and setting parameters of a medical instrument provided in an embodiment of the present invention;

[0019] Figure 2It is a result diagram of Frideman test on five optimization problem models based on IGD index in the embodiment of the present invention and other methods in the prior art;

[0020] Figure 3 It is a result diagram of Frideman test on five optimization problem models based on HV index according to the embodiment of the present invention and other methods in the prior art. DETAILED DESCRIPTION

[0021] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings and embodiments, but the implementation manner of the present invention is not limited thereto.

[0022] Example

[0023] This embodiment provides a method for intelligent optimization and setting of medical instrument parameters based on an evolutionary algorithm, which can be applied to the parameter optimization and setting process of medical instruments such as laser cutting instruments and high-frequency electric knives. Figure 1 As shown, the intelligent optimization setting method of this embodiment mainly includes the following steps:

[0024] S1. Analyze the medical instrument parameter optimization problem, determine the objective function and constraints, and establish a mathematical model for the optimization problem.

[0025] During the specific implementation, the medical industry field knowledge related to gastrointestinal polypectomy, such as the parameters, usage methods, and usage rules of medical instruments such as laser cutting instruments and high-frequency electrosurgical units, can be collected to form constraints; and a mathematical model of the optimization problem can be established with the optimization goals of maximizing cutting accuracy and minimizing damage to surrounding healthy tissues.

[0026] S2. Initialize the population of medical instrument parameters to be optimized to form a parent population P with a total of NP individuals.

[0027] In specific implementation, the population can be initialized by random initialization to form the parent population P. The formula for random initialization is as follows:

[0028]

[0029] Among them, X i,j represents the jth dimension of the ith individual, D is the number of medical instrument parameters to be optimized, X j and denote the lower and upper limits of the j-th dimension, respectively. is in the interval The random number is uniformly distributed on the surface, and in this embodiment, NP=100.

[0030] S3. Based on the crossover operator, generate an offspring population oP with a total of NP individuals; based on the mutation operator, perform mutation operation on the NP individuals of the offspring population to obtain the mutated offspring population; merge the parent population P with the mutated offspring population to form a new population Q.

[0031] In specific implementation, the steps of generating offspring individuals based on the crossover operator include:

[0032] S311, randomly select two individuals X1 and X2 from the parent population, and calculate the mean μ of the two individuals in each dimension j j and variance

[0033] S312, generate a random number ζ uniformly distributed between [0,1] q , combined with the parameters controlling the distribution index, the amplification factor β is calculated according to the following formula:

[0034]

[0035] Among them, η q To control the distribution index parameter, the value of this embodiment is 2.

[0036] S313, based on the individual mean and variance calculated in each dimension, and the parameters controlling the distribution index, two offspring individuals C1 and C2 are generated according to the following formula:

[0037]

[0038] S314. Repeat steps S311 to S313 until NP offspring individuals are generated.

[0039] In specific implementation, the steps of offspring individual mutation based on the mutation operator include:

[0040] S321, generate a random number ζ uniformly distributed between [0,1] m .

[0041] S322, calculate the variation θ according to the following formula j :

[0042]

[0043] Among them, β m It is a parameter for controlling the degree of variation, and its value in this embodiment is 20.

[0044] S323, for each individual X in the offspring population i , according to the following formula, the mutation operation is performed with a probability of 0.1 to obtain the mutant individual X i ′:

[0045]

[0046] Here, ζ is a random number between [0,1].

[0047] S324. Repeat steps S321 to S323 until mutation operations are performed on NP individuals of the offspring population.

[0048] S4. Based on the self-balancing fitness function, the fitness of each individual in the new population Q is calculated, and NP individuals with the smallest fitness are selected to form a new parent population.

[0049] In specific implementation, a new parent population can be formed through the following specific process:

[0050] S41, calculate the X of each individual in the new population Q according to the following formula: i Feasibility of F cv,i (X i );

[0051]

[0052] Among them, g j (X i ) represents the jth inequality constraint in the mathematical model of the optimization problem, h j (X i ) represents the jth equality constraint in the mathematical model of the optimization problem, δ represents the relaxation factor of the equality constraint, k represents the number of inequality constraints in the mathematical model of the optimization problem, and Lk represents the number of equality constraints in the mathematical model of the optimization problem.

[0053] S42, calculate the X of each individual in the new population Q according to the following formula: i The convergence of F pr,i (X i ):

[0054]

[0055] Among them, f j,min represents the minimum value of the jth objective function in the mathematical model of the optimization problem, f j,max It represents the maximum value of the jth objective function in the mathematical model of the optimization problem, and m represents the number of objective functions in the optimization problem.

[0056] S43, calculate the X of each individual in the new population Q according to the following formula: i Diversity F cd,i (X i ):

[0057]

[0058]

[0059] Among them, Y represents the parent population P except X i For individuals other than i ,Y) represents individual X i The Euclidean distance between X and Y; sh(X i ,Y) represents individual X i Shared function with Y; rand() is a random number, randomly selected from and Select from; r is a factor that depends on the population size and the number of objective functions, m represents the number of objective functions in the optimization problem.

[0060] S44, combining the feasibility, diversity and convergence of each individual in the new population, and calculating the fitness F of each individual in the new population Q based on the following self-balancing fitness calculation formula: i :

[0061]

[0062] Wherein, ω is the coefficient of convergence. In specific implementation, ω=5; α=0.5+0.5*p f is the adaptive balance coefficient, p f is the proportion of feasible solutions in the parent population P when the sub-band individuals are not integrated.

[0063] S5. Repeat step S3 and step S4 until the termination condition is reached, and a set of Pareto optimal solutions, that is, a set of optimal medical instrument parameters, is obtained.

[0064] In a specific implementation, the termination condition may be a preset number of iterations. When performing iterations, steps S3 to S4 may be repeated until the training reaches the preset number of times to obtain the final optimal medical instrument parameters.

[0065] The intelligent optimization setting method for medical instrument parameters based on evolutionary algorithm provided in this embodiment, through the proposed self-balancing fitness function, adaptively balances feasibility and diversity in the infeasible stage; in the semi-feasible stage, shifts from the balance between feasibility and diversity to the balance between convergence and diversity; in the feasible stage, adaptively balances convergence and diversity. Therefore, this embodiment can adapt to different mathematical models of optimization problems and has strong applicability. In addition, the optimization algorithm of this embodiment not only has fewer parameters, but also can ensure the feasibility and accuracy of the optimal solution, thereby ensuring the quality and safety of the operation.

[0066] Since the parameters of medical instruments in the mathematical model of the optimization problem actually used involve privacy protection, in the embodiment of the present invention, five widely used public optimization problems are taken as examples (as shown in Table 1) to replace the mathematical model of the optimization problem actually used to verify the effectiveness and superiority of the method proposed in this embodiment.

[0067] Table 1

[0068]

[0069]

[0070]

[0071] The five optimization problems in Table 1 are solved by using the method of this embodiment and the comparative method in the prior art, and the performance of each method is measured by the IGD index and the HV index. Both the IGD index and the HV index are indicators for measuring the comprehensive performance of the multi-objective optimization method. The smaller the IGD index, the better the performance of the corresponding method, and the larger the HV index, the better the performance of the corresponding method.

[0072] Each method was run independently 25 times, and its mean and standard deviation were calculated. The experimental results are shown in Tables 2 and 3, where "+" indicates that the comparison method is better than the method of this embodiment, "≈" indicates that the comparison method and the method of this embodiment have similar performance, and "-" indicates that the comparison method is inferior to the method of this embodiment. The experimental results are displayed in the form of "mean (standard deviation)". It can be seen from Table 2 that for the IGD index, the method of this embodiment is significantly better than the comparison method in 5, 4, 4 and 5 optimization problems, respectively. It can be seen from Table 3 that for the HV index, the method of this embodiment is significantly better than the comparison method in 5, 3, 3 and 5 optimization problems, respectively. The above experiments demonstrate the superiority and effectiveness of the method of this embodiment, and also verify the applicability of the method of this embodiment to different optimization problems.

[0073] At the same time, in order to further illustrate the comprehensive performance of each method on the five optimization problems, based on the Frideman test, statistical analysis is performed on the method of this embodiment and the comparative method on different indicators. The analysis results are as follows: Figure 2 and Figure 3 As shown. Figure 2 and Figure 3 It can be seen that the methods in this embodiment have achieved the minimum score, which proves that the methods in this embodiment are significantly better than the comparison methods and have statistical significance.

[0074] Table 2

[0075]

[0076] Table 3

[0077]

[0078] This embodiment also provides a storage medium, on which computer instructions are stored. When the computer instructions are executed by a processor, steps S1-S5 of the above-mentioned intelligent optimization setting method are implemented.

[0079] In addition, this embodiment also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor runs the computer program, steps S1-S5 of the above-mentioned intelligent optimization setting method are executed.

[0080] Although the embodiments of the present invention are described above in conjunction with the accompanying drawings, the present invention is not limited to the specific embodiments, which are merely illustrative and not restrictive. All other embodiments obtained by a person skilled in the art under the guidance of the present invention without departing from the scope of protection of the present invention and the claims shall fall within the scope of protection of the present invention.

Claims

1. A method for intelligent optimization setting of medical instrument parameters based on evolutionary algorithm, characterized in that: The following steps are involved: S1. Analyze the medical instrument parameter optimization problem, determine the objective function and constraints, and establish a mathematical model for the optimization problem; S2, initializing the population of medical instrument parameters to be optimized to form a parent population P with a total of NP individuals; S3. Based on the crossover operator, a total of NP individuals in the offspring population oP is generated; based on the mutation operator, the NP individuals in the offspring population are mutated to obtain the mutated offspring population; the parent population P is merged with the mutated offspring population to form a new population Q; S4. Based on the self-balancing fitness function, the fitness of each individual in the new population Q is calculated, and NP individuals with the smallest fitness are selected to form a new parent population; S5. Repeat steps S3 and S4 until a termination condition is reached to obtain a set of optimal medical instrument parameters.

2. The intelligent optimization setting method according to claim 1, characterized in that: Step S2 uses random initialization to initialize the population. The formula for random initialization is: Among them, X i,j represents the jth dimension of the ith individual, D is the number of medical instrument parameters to be optimized, X j and denote the lower and upper limits of the j-th dimension, respectively. is in the interval A uniformly distributed random number.

3. The intelligent optimization setting method according to claim 1, characterized in that: Step S3: The step of generating offspring individuals based on the crossover operator includes: S311, randomly select two individuals X1 and X2 from the parent population, and calculate the mean μ of the two individuals in each dimension j j and variance S312, generate a random number ζ uniformly distributed between [0,1] q , combined with the parameters controlling the distribution index, calculate the amplification factor β; S313, generating two offspring individuals based on the individual mean and variance calculated in each dimension, and the parameter controlling the distribution index; S314. Repeat steps S311 to S313 until NP offspring individuals are generated.

4. The intelligent optimization setting method according to claim 3, characterized in that: The formula for calculating the amplification factor β in step S312 is: Among them, η q is the parameter that controls the distribution index; In step S313, two offspring individuals are generated using the following formula:

5. The intelligent optimization setting method according to claim 2, characterized in that: Step S3: The offspring individual mutation step based on the mutation operator includes: S321, generate a random number ζ uniformly distributed between [0,1] m ; S322. Calculate the variation θ according to the following formula j : Among them, β m is a parameter that controls the degree of variation; S323, for each individual X in the offspring population i , according to the following formula, the mutation operation is performed with a probability of 0.1 to obtain the mutant individual X i ′: Among them, ζ is a random number between [0,1]; S324. Repeat steps S321 to S323 until mutation operations are performed on NP individuals of the offspring population.

6. The intelligent optimization setting method according to claim 1, characterized in that: Step S4: The steps of forming a new parent population include: S41. Calculate the number of each individual X in the new population Q i Feasibility of F cv,i (X i ); Among them, g j (X i ) represents the jth inequality constraint in the mathematical model of the optimization problem, h j (X i ) represents the jth equality constraint in the mathematical model of the optimization problem, δ represents the relaxation factor of the equality constraint, k represents the number of inequality constraints in the mathematical model of the optimization problem, and Lk represents the number of equality constraints in the mathematical model of the optimization problem; S42, calculate each individual X in the new population Q i The convergence of F pr,i (X i ): Among them, f j,min represents the minimum value of the jth objective function in the mathematical model of the optimization problem, f j,max represents the maximum value of the jth objective function in the mathematical model of the optimization problem, and m represents the number of objective functions in the optimization problem; S43, calculate each individual X in the new population Q i Diversity F cd,i (X i ): Among them, Y represents the parent population P except X i For individuals other than i ,Y) represents individual X i The Euclidean distance between X and Y; sh(X i ,Y) represents individual X i Shared function with Y; rand() is a random number, randomly selected from and Select from; r is a factor that depends on the population size and the number of objective functions; S44. Combine the feasibility, diversity and convergence of each individual in the new population, and calculate the fitness F of each individual in the new population Q based on the self-balancing fitness calculation formula. i : Where ω is the coefficient of convergence; α=0.5+0.5*p f is the adaptive balance coefficient, p f is the proportion of feasible solutions in the parent population P when the sub-band individuals are not integrated.

7. A storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed by the processor, each step of the intelligent optimization setting method described in any one of claims 1 to 6 is implemented.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor runs the computer program, the processor executes the steps of the intelligent optimization setting method according to any one of claims 1 to 6.

Images