Optimization Method for Main Girder of Bridge Erection Machine Based on Golden Sine Hybrid Fruit Fly Optimization Algorithm

By dividing the main beam into superior and inferior subgroups in the fruit fly optimization algorithm and combining the position update strategy of the golden sine and Chebyshev mapping, the problem of heavy self-weight in the design of the main beam of the bridge erecting machine is solved, realizing the lightweighting and efficient optimization of the main beam, and improving the accuracy and efficiency of the algorithm.

CN117610134BActive Publication Date: 2026-11-17CHINA RAILWAY 12TH BUREAU GRP CO LTD +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202311658082.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-06
Publication Date
2026-11-17
Estimated Expiration
2043-12-06

AI Technical Summary

Technical Problem

The existing design of the main beam of the bridge erecting machine mainly relies on experience and analogy, resulting in heavy weight, high material consumption, and the fruit fly optimization algorithm is prone to getting trapped in local optima, making it difficult to achieve efficient optimization.

Method used

A golden sine hybrid fruit fly optimization algorithm is adopted to divide the population into superior and inferior subgroups. The golden sine algorithm and Chebyshev mapping are used for position updates. The location of food sources is dynamically updated by combining neighborhood centroid back learning and Cauchy perturbation strategy, thereby improving the accuracy and efficiency of the algorithm.

Benefits of technology

It improved the accuracy and efficiency of the main beam optimization of the bridge erecting machine, realized the lightweight design of the main beam, reduced its self-weight and material consumption, and avoided the problems of premature convergence of the algorithm and getting trapped in local optima. The optimization rate reached 3.64%.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117610134B_ABST
    Figure CN117610134B_ABST
Patent Text Reader

Abstract

This invention belongs to the field of building construction and discloses a method for optimizing the main beam of a bridge erecting machine based on the Golden Sine Hybrid Fruit Fly Optimization Algorithm. The method includes the following steps: Step 1: The cross-sectional area of ​​the main beam is the optimization target; the objective function and corresponding key parameters to be optimized are determined. Step 2: The Golden Sine Hybrid Fruit Fly Optimization Algorithm optimizes each key parameter. This algorithm divides the population into a superior subgroup and a inferior subgroup. The position of individuals in the superior subgroup is updated using the Golden Sine algorithm. In the inferior subgroup, the position of individual fruit flies is updated using chaotic mapping. Then, the food source position is dynamically updated using a hybrid mutation strategy combining neighborhood centroid back learning and Cauchy perturbation. Step 3: Based on the optimization results of the key parameters, the strength and stiffness of the main beam structure are checked and calculated. This invention can improve the optimization effect and efficiency of the main beam of a bridge erecting machine.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of building construction, specifically relating to an optimization method for the main beam of a bridge erecting machine based on the Golden Sine Algorithm of Fruit Fly Optimization Algorithm (GSAFOA). Background Technology

[0002] With the continuous expansion of domestic infrastructure and transportation construction, and the actual demand for improving travel conditions and efficiency, the construction conditions of high-speed rail are becoming increasingly complex. Challenges include small curves, complex terrain, and severely limited construction space, necessitating the erection of precast box girders. Bridge erecting machines are key equipment for this purpose, making their importance in the construction process self-evident. A bridge erecting machine mainly consists of a metal structure, working mechanism, and electrical control equipment. The metal structure bears the weight of the entire machine and the load, while the main girder, as the most critical component, typically accounts for 60% to 80% of the machine's weight. Currently, the design of bridge erecting machine series products is primarily based on experience and analogy. The main girder structure designed using traditional methods, due to its large safety margin, results in a heavy machine and high material consumption. To save resources and reduce costs, it is necessary to optimize the structure of the bridge erecting machine by reducing its weight while ensuring construction safety and meeting design requirements for mechanical performance.

[0003] The Fruit Fly Optimization Algorithm (FOA) is an intelligent biomimetic algorithm that seeks global optimization based on the foraging behavior of fruit flies. Fruit flies possess superior olfactory and visual senses compared to other species. Their olfactory organs can collect various odors floating in the air, determining potential food sources based on the optimal odor concentration. When a fruit fly flies near a food location, its keen vision identifies the location of food and its companions, and it flies in that direction until it finds food. However, the FOA learns only from the current best individual during the iterative optimization process. Once the best individual for the current iteration is found, all individuals gather at that individual's location, reducing population diversity and making the algorithm prone to getting trapped in local optima. This leads to lower convergence speed and accuracy, resulting in premature convergence. Summary of the Invention

[0004] This invention overcomes the shortcomings of existing technologies and aims to solve the following technical problem: providing a method for optimizing the main beam of a bridge erecting machine based on a hybrid fruit fly optimization algorithm with a golden sine curve. The fruit fly algorithm is improved by dividing the overall population into superior and inferior subgroups. The positions of the superior and inferior subgroups are updated using the golden sine curve algorithm and Chebyshev mapping, respectively, thereby improving the solution accuracy and efficiency of the fruit fly algorithm, and thus realizing the optimization of the main beam of the bridge erecting machine and the lightweight design of the main beam.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: a method for optimizing the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm, comprising the following steps:

[0006] Step 1: Using the cross-sectional area of ​​the main beam as the optimization objective, determine the objective function and the corresponding key parameters to be optimized;

[0007] Step 2: Optimize the key parameters using the Golden Sine Hybrid Fruit Fly Optimization Algorithm, specifically including the following steps:

[0008] Step 2.1: Initialize parameters;

[0009] Step 2.2: Initialize the search direction and distance of the fruit flies, perform basic fruit fly transformations, and calculate the fitness of each individual fruit fly;

[0010] Step 2.3: Based on the fitness of each individual fruit fly, determine the best and worst fruit fly individuals and their coordinates in the current fruit fly population;

[0011] Step 2.4: Based on the coordinates of the best and worst fruit fly individuals, calculate the distance D between each fruit fly individual in the population and the best and worst fruit fly individuals, respectively. i,b and D i,w ;

[0012] Step 2.5: If D i,b ≤D i,w Then, the fruit fly individuals are assigned to a better subgroup, and the positions of individuals in the better subgroup are updated using the golden sine algorithm; if D i,b >D i,w If the corresponding fruit fly individual is assigned to the inferior subgroup, the position of the fruit fly individual in the inferior subgroup is updated by chaotic mapping.

[0013] The formula for updating the position of individuals in the superior subgroup using the golden sine algorithm is as follows:

[0014]

[0015] In the formula, P represents the position of individual fruit fly i (i = 1, 2, ..., s) at the t-th iteration.i t The optimal position of fruit fly individual i in the tth generation selection, R1 is a random number in the interval [0, 2π], which determines the moving distance of fruit fly individual i in the next iteration, R2 is a random number in the interval [0, 2π], which controls the moving direction of individual i in the next iteration, m1 and m2 are coefficients obtained based on the golden ratio t, m1 = -π + (1 - t) * 2π, m2 = -π + t * 2π;

[0016] Step 2.6: Merge the better and worse subgroups, calculate the fitness of each fruit fly individual at its new position, and record the fruit fly individual with the best fitness in the fruit fly population and its position.

[0017] Step 2.7: Dynamically update the food source location through hybrid mutation, calculated using the following formula:

[0018]

[0019] Among them, F new Let be the new location of the food source after mixing and mutation, k represent the contraction factor, which is a random number uniformly distributed in the interval [0,1], F is the optimal food source location in the current iteration, cauchy(0,1) represents the Cauchy probability objective function, and R3 represents a random number uniformly distributed in the interval [0,1].

[0020] Step 2.8: Recalculate the fitness of each fruit fly individual based on the updated food source location; record the optimal fitness value and the coordinates of the corresponding fruit fly individual, and make the fruit fly swarm fly towards that location using visual advantage;

[0021] Step 2.9: Determine if the maximum number of iterations has been reached. If not, return to step 2.2 and repeat steps 2.2 to 2.8 until the iteration ends, then output the optimal solution.

[0022] Step 3: Based on the optimization results of the key parameters, perform verification calculations on the strength and stiffness of the main beam structure.

[0023] In step 1, the expression for the objective function is:

[0024] f = x1x2 + x3x6 + x3x7 + x4x5;

[0025] Where x1 is the width of the upper flange, x2 is the thickness of the upper flange, x3 is the height of the web, x4 is the thickness of the lower flange, x5 is the width of the lower flange, x6 and x7 are the thicknesses of the webs on both sides, and x8 is the distance between the webs; the key parameters to be optimized include x1 to x8.

[0026] In steps 2.2 and 2.5, the specific formulas for performing the basic fruit fly transformation are as follows:

[0027] X i =X_axis + RandomValue

[0028] Y i =Y_axis + RandomValue

[0029] Among them, X i and Y i X_axis and Y_axis represent the coordinates of the i-th fruit fly individual after the basic fruit fly transformation, respectively; X_axis and Y_axis represent the coordinates of the i-th fruit fly individual before the basic fruit fly transformation, respectively; and RandomValue represents the search distance.

[0030] In step 2.4, the distance D between each individual fruit fly in the population and the best and worst fruit fly individuals is... i,b and D i,w The calculation formula is:

[0031]

[0032]

[0033] Among them, D i,b and D i,w X represents the distance between the i-th fruit fly individual and the best and worst fruit fly individuals, respectively; b Y b X represents the coordinates of the optimal fruit fly individual. w and Y w This represents the coordinates of the worst-performing fruit fly individual.

[0034] In step 2.2, the fitness of each individual fruit fly is calculated as follows:

[0035] First, calculate the distance between the individual fruit fly and the origin. Then, calculate the individual's taste concentration judgment value using the following formula:

[0036]

[0037]

[0038] Among them, Dist i S represents the distance between the i-th fruit fly individual and the origin. i This represents the taste concentration judgment value of the i-th fruit fly individual;

[0039] Then, the taste concentration determination value S i Substitute the flavor concentration into the flavor concentration determination function to calculate the flavor concentration of each individual fruit fly as its fitness.

[0040] In step 2.3, the method for determining the best and worst fruit fly individuals in the current fruit fly population is as follows:

[0041] [bestSmell bestIndex1]=min(Smell i )

[0042] [worseSmell bestIndex2]=max(Smell i )

[0043] Where [bestSmell bestIndex1] represents the best fruit fly individual, [worseSmell bestIndex2] represents the worst fruit fly individual, and Smell... i This represents the fitness value of the i-th fruit fly individual.

[0044] Step 2.8 specifically includes the following steps:

[0045] Determine whether the optimal fitness value obtained after updating the food source location through mixed mutation is less than the optimal fitness value of the fruit fly population obtained before the update. If yes, update the food source location; otherwise, do not update and retain the original food source location information.

[0046] Then, the optimal fitness value and its corresponding coordinates of the fruit fly individuals are recorded and preserved, and the fruit fly swarm is made to fly to that location using vision.

[0047] In step 2.1, the initialized parameters include population size Sizepop, maximum number of iterations Maxgen, fruit fly population location, and fitness variance threshold δ.

[0048] In step 2.5, the formula for updating the position of individual fruit flies in the superior subgroup using chaotic mapping is as follows:

[0049]

[0050]

[0051] CX′ i =4CX i *(1-CX i )

[0052] CY′ i =4CY i *(1-CY i )

[0053] X′ i =a i +CX′ i (b i -ai )

[0054] Y′ i =a i +CY′ i (b i -a i )

[0055] Among them, a i and b i Let CX represent the control parameters of the i-th variable, respectively. i CY i This indicates the position X of the individual fruit fly. i Y i The chaotic variable CX obtained by performing chaotic mapping i 'and CY i 'Indicates according to CX i CY i The intermediate variable obtained, X i ', Y i 'Indicates the new location of a fruit fly individual within the search space.

[0056] In step 3, the specific method for verifying the strength and stiffness of the main beam structure is as follows:

[0057] Calculate the deformation, strength, and static stiffness under working conditions one and two, and determine whether they meet the design requirements; working condition one refers to the lifting trolley being located at both ends of the main beam; working condition two refers to the crane being at the mid-span position.

[0058] Compared with the prior art, the present invention has the following advantages:

[0059] 1. This invention provides a method for optimizing the main beam of a bridge erecting machine based on a hybrid fruit fly optimization algorithm with a golden sine curve. The method divides the overall population into superior and inferior subgroups based on the fruit fly optimization algorithm. The golden sine curve algorithm and Chebyshev mapping are used to update the positions of the superior and inferior subgroups respectively. The two subgroups are then merged, and the fitness of the fruit flies in the merged population is evaluated. The position of the best-fit fruit fly in the merged population is then dynamically updated by alternately applying a neighborhood centroid back-learning strategy and a Cauchy perturbation strategy with equal probability. The neighborhood centroid back-learning strategy expands the search range of the algorithm, enabling it to discover more potential global optima in the search space, thus improving the algorithm's optimization accuracy. The random perturbation of the food sources using the Cauchy mutation operator enhances population diversity to a certain extent and reduces the probability of the algorithm getting trapped in local optima, thereby improving the algorithm's accuracy.

[0060] 2. After the hybrid mutation is completed, the present invention will decide whether to update the food source location based on the greedy principle. That is, the food source location will only be updated when the fitness of the mutated individual is better than that of the original individual. Otherwise, the original food source location information will be retained. Finally, the optimal solution of concentration will be obtained based on the location information, which further improves the optimization accuracy.

[0061] 3. Experiments have shown that the improved algorithm of this invention outperforms existing fruit fly optimization algorithms and the Dual Strategy Evolutionary of Fruit Fly Optimization Algorithm (DSEFOA) in terms of minimum, maximum, average, and variance of the optimal solutions under five standard test functions. This indicates that the optimal solutions obtained by the algorithm of this invention are more stable, have lower dispersion, and exhibit faster convergence speed and convergence.

[0062] 4. The GSAFOA algorithm used in this invention achieves the highest optimization rate in the iterative optimization of the main beam cross-section size of the bridge erecting machine compared to FOA and DSEFOA. It avoids the problems of premature convergence and getting trapped in local optima in the FOA algorithm, and the weak global search capability of the DSEFOA algorithm, which cannot significantly improve the accuracy of the optimization results compared to the FOA algorithm. The GSAFOA algorithm used in this invention has better global and local search capabilities. During the iterative optimization process, this algorithm can continuously escape local optima and converge to a high-precision solution. When the population size is set to 50 and the number of iterations is set to 500, the cross-section size of the main beam of the SXJ900 bridge erecting machine optimized by this invention is reduced from the original 115680mm. 2 Reduced to 111474mm 2 The optimization rate reached 3.64%, demonstrating significant optimization results. Attached Figure Description

[0063] Figure 1 This is a flowchart illustrating the golden sine hybrid fruit fly optimization algorithm used in an embodiment of the present invention.

[0064] Figure 2 The images show the top view (a) and mid-span section (b) of the main beam used in this embodiment of the invention.

[0065] Figure 3 A simplified diagram of the stress on the main beam of the bridge erecting machine;

[0066] Figure 4 A simplified diagram showing the stress on one end of a concrete box girder when it is lifted to the mid-span of the main girder.

[0067] Figure 5 This is a three-dimensional image diagram of the five standard test functions used in the embodiments of the present invention;

[0068] Figure 6 The graph shows the iterative optimization curves obtained based on five standard test functions.

[0069] Figure 7 The diagram shows the iterative calculation for optimizing the cross-sectional dimensions of the main beam of the bridge erecting machine. Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0071] This invention provides a method for optimizing the main beam of a bridge erecting machine based on a golden sine hybrid fruit fly optimization algorithm. Figure 1 The method shown here is a lightweight design of the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm, which specifically includes the following steps:

[0072] Step 1: Using the cross-sectional area of ​​the main beam as the optimization target, determine the objective function and the corresponding key parameters to be optimized.

[0073] The main girder structure of the bridge erecting machine consists of box girders, which are spliced ​​from steel plates. The goal of optimizing the metal structure is to reduce material consumption and overall equipment weight while meeting structural requirements, thereby improving economic efficiency. Taking the SXJ900 / 32 bridge erecting machine as an example, its main girder mid-span section is as follows... Figure 2 As shown, its cross-section optimization parameters are expressed by a vector formula:

[0074] X=[x1,x2,x3,x4,x5,x6,x7,x8] (1)

[0075] Where x1 is the width of the upper flange, x2 is the thickness of the upper flange, x3 is the height of the web, x4 is the thickness of the lower flange, x5 is the width of the lower flange, x6 and x7 are the thicknesses of the webs on both sides, and x8 is the distance between the webs. The key parameters of the mid-span section of the main beam are shown in Table 1.

[0076] Table 1 Key parameters of the mid-span section of the main beam

[0077]

[0078] For the lightweight design of the main beam section between the front and rear legs of the SXJ900 / 32 bridge erecting machine, the main beam section has the same cross-sectional dimensions. The goal is to minimize its weight, that is, to minimize the cross-sectional area of ​​the beam. The objective function is shown in the following formula (2).

[0079] f(x)=x1x2+x3x6+x3x7+x4x5 (2)

[0080] The material used is Q345, with an elastic modulus E of 206 GPa, a Poisson's ratio μ of 0.3, and a density of 7850 kg / m³. 3 Referring to GB / T 3811-2008 "Code for Design of Cranes", the safety factor is taken as n = 1.48, and the allowable stress [σ] is 233MPa. Referring to the "Crane Design Manual", the strength and stiffness of the main beam structure are calculated under two conditions of high stress for the bridge erecting machine, that is, the strength and stiffness of the main beam structure are taken as constraints.

[0081] Working Condition 1: When erecting a 32m beam (the 32m concrete beam is the heaviest, so we take 900t for analysis), the entire machine is supported by the front and rear outriggers. The front and rear overhead cranes lift the concrete beam. Figure 3 As shown, A and B are the outriggers of the bridge erecting machine, and the lifting trolley is located at both ends of the main beam. The two support points of the main beam are simplified to hinge points, and the lifting points of the lifting trolley are simplified to concentrated force applications. The self-weight load of the main beam of the bridge erecting machine is a uniformly distributed load q, F. j For the self-weight of hoist 2 and other components, the comprehensive influence load factor is taken as 1.1.

[0082] According to the Crane Design Manual, the bending moment of the main beam at mid-span in the vertical direction is shown in equation (3).

[0083]

[0084] The shear force at mid-span of the main beam in the vertical direction can be solved using the force and moment equilibrium equations:

[0085] ∑F=F1+F2-qL-F c1 -F c2 =0 (4)

[0086]

[0087] In the formula, F c1 F c2 Let q be the self-weight of the crane trolley and the lifting load (N), q be the uniformly distributed gravity load of the main beam (N), F1 and F2 be the support reaction forces of the two legs of the main beam (N), L be the length of the main beam (mm), l be the distance between the two crane trolleys (mm), and l' be the distance between the crane trolley and the support end B.

[0088] According to the strength measurement, the stress at the corner of the flange plate at mid-span section must meet the following requirements:

[0089]

[0090] In the formula, Mx M y W is the bending moment of the box girder section. x W y The section modulus of bending of the main beam section (N·mm).

[0091] The shear stress at the mid-span section must satisfy:

[0092]

[0093] In the formula, F is the shear force (N);

[0094] δ represents the web thickness (mm).

[0095] τ represents the shear stress at the mid-span section;

[0096] Sx represents the static moment along the x-axis;

[0097] [τ] represents the allowable shear stress of the material (material parameter of the main beam of the bridge erecting machine).

[0098] The deflection at mid-span of the main beam section is:

[0099]

[0100]

[0101] In the formula, f represents deflection;

[0102] [f] indicates the allowable deflection of the main girder (parameter of the main girder of the bridge erecting machine);

[0103] E is the elastic modulus (MPa);

[0104] I x Moment of inertia of the main beam section of the bridge erecting machine (mm) 4 ).

[0105] Working Condition 2: When erecting a 32m beam (the 32m concrete beam is the heaviest, so we take 900t for analysis), the entire machine is supported by the front and rear outriggers. Analysis shows that the front and rear overhead cranes lift the concrete beam. The front overhead crane (F...) c3 At the mid-span position (L / 2), the bending moment and flexural deformation of the main beam are at their maximum. The two supports of the main beam are simplified to hinge points, and the lifting points of the hoisting trolley are simplified to concentrated force. The self-weight load of the main beam of the bridge erecting machine is a uniformly distributed load q, F. j For the self-weight of hoist 2 and other components, the comprehensive influence load factor is taken as 1.1. The stress analysis is as follows: Figure 4 As shown.

[0106] The bending moment of the main beam at mid-span in the vertical direction is:

[0107]

[0108] Under this condition, the deflection is:

[0109]

[0110] Under both working conditions, the shear stress on the web should satisfy... Design requirements.

[0111] Q represents the calculated shear force (N);

[0112] I represents the moment of inertia of the calculated section (mm). 4 );

[0113] S i Let be the static moment of the web.

[0114] At the same time, according to the static stiffness criterion, That is, f 刚 ≤50.1mm.

[0115] Among them, f 刚 This indicates the static deflection under rated load.

[0116] The overall stability of the main beam is calculated as follows: the height-to-width ratio of the main beam, h / b, is (2960+16+20) / (1048+12+12) = 2.8 < 3. According to the design specification for cranes, GB3811-2008, the requirement is met.

[0117] Step 2: Optimize the key parameters using the Golden Sine Hybrid Fruit Fly Optimization Algorithm, specifically including the following steps:

[0118] Step 2.1: Initialize parameters.

[0119] In step 2.1, the initialized parameters include population size Sizepop, maximum number of iterations Maxgen, fruit fly population location, and fitness variance threshold δ.

[0120] Step 2.2: Initialize the search direction and distance of the fruit flies, perform basic fruit fly transformations, and calculate the fitness of each individual fruit fly.

[0121] In step 2.2, the basic fruit fly transformation is first performed, and the specific formula is as follows:

[0122] X i =X_axis + RandomValue (12)

[0123] Y i =Y_axis + RandomValue (13)

[0124] Among them, X i and Y iX_axis and Y_axis represent the coordinates of the i-th fruit fly individual after the basic fruit fly transformation, respectively. RandomValue represents the search distance, which is a random value.

[0125] Since the location of the food is unknown, the distance between the individual fruit fly and the origin (Dist) is estimated. i Then calculate the individual fruit fly taste concentration judgment value S. i This value is the reciprocal of the distance:

[0126]

[0127]

[0128] Finally, the taste concentration determination value S was determined. i Substituting the values ​​into the taste concentration determination function, we can calculate the taste concentration (Smell) of an individual fruit fly. i That is, fitness.

[0129] Smell i =Function(S i (16)

[0130] Wherein, Function represents the flavor concentration determination function.

[0131] Step 2.3: Based on the fitness of each individual fruit fly, determine the best and worst fruit fly individuals and their coordinates in the current fruit fly population.

[0132] In step 2.3, the method for determining the best and worst fruit fly individuals in the current fruit fly population is as follows:

[0133] [bestSmell bestIndex1]=min(Smell i (17)

[0134] [worseSmell bestIndex2]=max(Smell i (18)

[0135] Where [bestSmell bestIndex1] represents the best fruit fly individual, [worseSmell bestIndex2] represents the worst fruit fly individual, and Smell... i This represents the fitness value of the i-th fruit fly individual.

[0136] Based on pointers bestIndex1 and bestIndex2, extract the best and worst fruit fly individual coordinates from the population coordinates Xi and Yi;

[0137]

[0138]

[0139] Step 2.4: Based on the coordinates of the best and worst fruit fly individuals, calculate the distance D between each fruit fly individual in the population and the best and worst fruit fly individuals, respectively. i,b and D i,w The calculation formula is:

[0140]

[0141] Among them, D i,b and D i,w X represents the distance between the i-th fruit fly individual and the best and worst fruit fly individuals, respectively; b Y b X represents the coordinates of the optimal fruit fly individual. w and Y w This represents the coordinates of the worst-performing fruit fly individual.

[0142] Step 2.5: If D i,b ≤D i,w Then, the fruit fly individuals are assigned to a better subgroup, and the positions of individuals in the better subgroup are updated using the golden sine algorithm; if D i,b >D i,w If the corresponding fruit fly individual is assigned to a less desirable subgroup, the position of the fruit fly individual in the less desirable subgroup is updated using chaotic mapping.

[0143] In step 2.5, the formula for updating the position of individuals in the superior subgroup using the golden sine algorithm is as follows:

[0144]

[0145] In the formula, This represents the position of fruit fly individual i (i = 1, 2, ..., s) at the t-th iteration, where d is the problem dimension. The optimal position of fruit fly individual i in the t-th generation is determined by R1, a random number in the interval [0, 2π], which determines the moving distance of fruit fly individual i in the next iteration. R2, a random number in the interval [0, 2π], controls the moving direction of individual i in the next iteration. m1 and m2 are based on the golden ratio. The obtained coefficients are m1 = -π + (1 - t) * 2π and m2 = -π + t * 2π. These coefficients narrow the algorithm's search space, guiding individuals to gradually approach the optimal value. The Golden Sine Algorithm is not designed to simulate natural phenomena, but rather is a mathematical model that uses the sine function and the golden ratio for iterative optimization, and the algorithm has good portability.

[0146] In step 2.5, the position X of the individual fruit fly is determined. i Y i Mapped to a chaotic variable CX using the Chebyshev method. i CY i Then, equations (25) and (26) are used to analyze the chaotic variable CX. i CY i The intermediate variable CX obtained by transformation i 'and CY i Finally, equations (27) and (28) are used to transform the intermediate variable CX. i 'and CY i 'Transformed into a new location X of a fruit fly individual within the search space' i ,Y′ i The specific formula for updating the location is:

[0147]

[0148]

[0149] CX i =4CX i *(1-CX i (25)

[0150] CY i =4CY i *(1-CY i (26)

[0151] X′ i =a i +CX i (b i -a i (27)

[0152] Y′ i =a i +CY i (b i -a i (28)

[0153] Among them, a i and b i Let CX be the control parameter of the i-th variable. i CY i This indicates the position X of the individual fruit fly. i Y i The chaotic variable CX obtained by performing chaotic mapping i 'and CY i 'Indicates according to CX i CY iThe intermediate variable obtained, X i ', Y i 'Indicates the new location of a fruit fly individual within the search space.

[0154] Step 2.6: Merge the better and worse subgroups, calculate the fitness of each fruit fly individual at its new position according to formulas (14) to (16), and record the fruit fly individual with the best fitness in the fruit fly population and its position.

[0155] Specifically, first use equation (14) to calculate the distance of the fruit fly individual from the origin when searching for food by smell, and then calculate the taste concentration judgment value of the new location according to equation (15); substitute the taste concentration judgment value into the taste concentration judgment function (16) to find the taste concentration of the fruit fly individual's location; then, find the fruit fly with the best taste concentration in the fruit fly group.

[0156] Step 2.7: Dynamically update the location of food sources through hybrid mutation.

[0157] This invention integrates a hybrid mutation strategy that combines neighborhood centroid reverse learning and Cauchy perturbation to randomly perturb the current optimal individual position, thereby enhancing population diversity and improving the algorithm's ability to escape local optima.

[0158] (1) Domain-centric reverse learning

[0159] Backward learning is a mathematical model proposed by Tizhoosh. Its main idea is to select a better solution by simultaneously evaluating the current solution and the backward solution. Traditional backward learning obtains a high-quality solution by calculating the backward point through the maximum and minimum boundaries. However, this method does not take into account the information exchange between populations. To address this, Rahnamayan et al. proposed a backward point based on the centroid.

[0160] To better guide individuals in finding the best, this invention incorporates neighborhood centroid back learning into the algorithm.

[0161] Let the optimal solution (food source) position in the current iteration be F(F=(F1,F2,F3,...,F...). D )), F1……F D Let M be the centroid of the neighborhood, representing the centroid of the entire fruit fly population. Then, the solution obtained after back-learning from the neighborhood centroid is... The calculation method is as follows:

[0162]

[0163] Where k is the contraction factor, which is a random number uniformly distributed between [0,1].

[0164] (2) Cauchy mutation

[0165] The Cauchy distribution is a continuous probability distribution with no expected value. A random variable x is said to follow a Cauchy distribution if its probability density function is satisfied. The probability density function of the standard Cauchy distribution is as follows:

[0166]

[0167] The Cauchy distribution is characterized by high probability density at the origin, compact distribution, low density at both ends, and a relatively long distribution. Cauchy mutation can generate greater perturbation to individuals, enabling them to escape local optima. The Cauchy mutation calculation is introduced into the algorithm to perturb the location of the food source, and the calculation method is as follows:

[0168] F'=F+F*cauchy(0,1) (31)

[0169] Where F' is the new position of the food source after Cauchy mutation, cauchy(0,1) represents the Cauchy probability objective function, and (0,1) represents taking an arbitrary number in the interval 0 to 1.

[0170] (3) Mixed variation

[0171] To enhance the algorithm's optimization capabilities, GSAFOA will alternately execute the neighborhood centroid back learning strategy and the Cauchy perturbation strategy with equal probability, dynamically updating the food source location. The hybrid mutation calculation method is as follows:

[0172]

[0173] Among them, F new Let represent the new location of the food source after mixing and mutation, k represents the contraction factor, which is a random number uniformly distributed in the interval [0,1], F is the optimal location of the food source in the current iteration, cauchy(0,1) represents the Cauchy probability objective function, and R3 represents a random number uniformly distributed in the interval [0,1].

[0174] In step 2.7, the reverse solution is generated by using the neighborhood centroid reverse learning strategy, which expands the search range of the algorithm and enables the algorithm to discover more potential global optimal solutions in the search space, thereby improving the algorithm's optimization accuracy. The Cauchy mutation operator is used to randomly perturb the food source, which enhances the population diversity to a certain extent and reduces the probability of the algorithm getting stuck in local optima.

[0175] Step 2.8: Substitute the updated food source location into the fitness function to recalculate the fitness of each fruit fly individual; record the optimal fitness value and the coordinates of the corresponding fruit fly individual, and make the fruit fly swarm fly to that location using visual advantage.

[0176] Furthermore, step 2.8 specifically includes the following steps:

[0177] Determine whether the optimal fitness value obtained after updating the food source location through mixed mutation is less than the optimal fitness value of the fruit fly population obtained before the update. If yes, update the food source location; otherwise, do not update and retain the original food source location information.

[0178] Then, the optimal fitness value and its corresponding coordinates of the fruit fly individuals are recorded and preserved, and the fruit fly swarm is made to fly to that location using vision.

[0179] Since step 2.7 cannot guarantee that the mutated individual is better than the original individual, in order to further improve the convergence accuracy of the algorithm, the GSAFOA algorithm of this embodiment will decide whether to update the food source position according to the greedy principle after the hybrid mutation is completed. That is, the food source position will only be updated when the fitness of the mutated individual is better than that of the original individual; otherwise, the original food source position information will be retained.

[0180] Step 2.9: Determine if the maximum number of iterations has been reached. If not, return to step 2.2 and repeat steps 2.2 to 2.8 until the iteration ends, then output the optimal solution.

[0181] Step 3: Based on the optimization results of the key parameters, perform verification calculations on the strength and stiffness of the main beam structure.

[0182] Specifically, in step 3, the method for verifying and calculating the strength and stiffness of the main beam structure is as follows:

[0183] Calculate the deformation, strength, and static stiffness under working conditions one and two, and determine whether they meet the design requirements; working condition one refers to the lifting trolley being located at both ends of the main beam; working condition two refers to the crane being at the mid-span position.

[0184] Algorithm test results and analysis:

[0185] (1) To test the characteristics of the fruit fly optimization algorithm and verify whether the optimization algorithm of this invention has an optimization effect on the basic fruit fly algorithm, five standard test functions f1 to f5 were used to perform performance tests on the FOA algorithm, the DSEFOA algorithm, and the GSAFOA algorithm of this invention. The test function expressions and parameter settings are shown in Table 2. The three-dimensional images of f1 to f5 are shown in Table 2. Figure 5As shown in (a) to (e), the FOA algorithm refers to the traditional fruit fly optimization algorithm, and the DSEFOA algorithm refers to the Dual Strategy Evolutionary of Fruit Fly Optimization Algorithm. Similar to the GSAFOA algorithm of this invention, it is based on the idea of ​​perturbing the overall population in the fruit fly algorithm. It divides the population by calculating the distance between an individual and the best individual. That is, based on the fruit fly optimization algorithm, the overall population is divided into superior and inferior subgroups. If the distance of individual i from the current best individual is less than or equal to the distance from the worst individual, then individual i is divided into a superior subgroup; otherwise, it is divided into a inferior subgroup. Unlike the GSAFOA algorithm of this invention, the two subgroups are processed using two different methods. The positions of individual fruit flies in the better subgroup are mapped to the chaotic interval [0,1] using Chebyshev. Then, the chaotic variables are transformed into the positions of individual fruit flies in the optimization search space. The inferior subgroup retains the original fruit fly processing method. Finally, the two subgroups are merged together, and then the position of the fruit fly with the best fitness is found.

[0186] Table 2 Standard Test Functions

[0187]

[0188] The initial parameters were set with a population size of 50, 500 iterations, and 15 runs. The results were statistically analyzed to obtain the minimum, maximum, average, and standard deviation of the 15 optimal solutions obtained by each algorithm. The test results are shown in Table 3. The convergence curves of the five standard test functions f1 to f5 are shown in the figure below. Figure 6 As shown in (a) to (e).

[0189] Table 3 Test Results

[0190]

[0191] In the test functions mentioned above, f5 is a unimodal function, which can be used to evaluate the algorithm's development capability, while f1, f2, f3, and f4 are multimodal functions, which can be used to evaluate the algorithm's search capability. As can be seen from Table 3 of the test results, when f1 to f5 are run 15 times using DSEFOA and GSAFOA, the minimum value obtained from the 15 runs is smaller than that obtained using FOA, and the average value of the optimal solutions obtained from the 15 runs is also smaller than the average value of the optimal solutions obtained using FOA. The difference is more pronounced with f1, where the average value of the optimal solutions obtained using FOA is 2.40 × 10⁻⁶. -3 The average value of the optimal solution using GSAFOA is 3.22 × 10⁻⁶. -67In addition, the average and maximum values ​​of the optimal solutions obtained by DSEFOA are larger than those obtained by GSAFOA. Therefore, DSEFOA's optimization ability needs improvement compared to GSAFOA. Standard deviation can reflect the stability of the algorithm to some extent. The test results show that the variances obtained by running f1 to f5 15 times using both DSEFOA and GSAFOA are smaller than those obtained using FOA, and GSAFOA has the smallest variance, indicating that the obtained optimal solutions are more stable and have lower dispersion. Figure 6 The algorithm iteration optimization curves show that the algorithm iteration graphs obtained using DSEFOA and GSAFOA have a faster convergence speed than FOA, and the minimum value reached in the algorithm iterations is also lower. This phenomenon is most obvious in the algorithm iteration graph of the f4 test function. Therefore, the two optimization algorithms have a faster convergence speed and stronger search ability compared to FOA. Furthermore, the analysis results above show that the curve obtained by GSAFOA in this invention has a higher slope, indicating that the algorithm has a faster convergence speed and good convergence performance.

[0192] (2) Optimization results of the main beam structure of the bridge erecting machine

[0193] For the optimization of various parameters of the bridge erecting machine, the population size was set to 50, the number of iterations was set to 500, and the two algorithms were run 30 times each. The optimal set (parameters X1-X8 were optimal, and the area was calculated to be optimal) was selected. The results are shown in Table 4. The optimization process is as follows: Figure 7 As shown, the optimized value of the calculation result represents the direct result obtained by the algorithm, and the rounded value represents the value considering the convenience of machining.

[0194] Table 4 Optimization Calculation Results of Each Parameter

[0195]

[0196]

[0197] As shown in Table 4, the GSAFOA algorithm achieved the highest optimization rate among the three algorithms. Although the DSEFOA algorithm performed well in the test function, its optimization results for the main beam cross-section size of the bridge erecting machine were similar to those of FOA. Figure 7In the analysis, the FOA algorithm converged prematurely, indicating that it may have gotten stuck in a local optimum, preventing it from exploring a better solution. While the DSEFOA algorithm did not converge prematurely, its global search capability was weak, and its accuracy in finding the optimal solution was not significantly improved compared to the FOA algorithm. The GSAFOA algorithm possessed better global and local search capabilities; during the iterative optimization process, it could continuously escape local optima and converge to a high-precision solution. The GSAFOA algorithm was used to optimize the main beam cross-section dimensions of the SXJ900 bridge erecting machine, reducing the original 115680mm... 2 Reduced to 111474mm 2 The optimization rate reached 3.64%, demonstrating significant optimization results.

[0198] Based on the optimization results of the key parameters of the main beam in Table 1, the strength and stiffness of the optimized main beam structure were checked and calculated using the aforementioned stress analysis and corresponding calculation formulas, and the strength design requirements were met. The static stiffness results show that the design requirement of a deflection-to-span ratio of 1 / 700 is met. The results are shown in Table 5 below.

[0199] Table 5. Calculation results of cross-sectional strength and stiffness for each algorithm.

[0200]

[0201] In summary, the main beam optimization method for bridge erecting machines of this invention is based on the dual-strategy fusion of the Golden Sine Hybrid Mutation Fruit Fly Algorithm (GSAFOA). Under the dual-strategy approach, the GSAFOA algorithm employs two methods to process the population: an adaptive chaotic algorithm and a golden sine leadership strategy. It also integrates a hybrid mutation algorithm combining neighborhood centroid back-learning and Cauchy perturbation strategies to update food source locations. The GSAFOA algorithm exhibits the strongest minimum search capability, high convergence speed and accuracy, and a low probability of getting trapped in local optima. Through testing with five benchmark functions, comparing the minimum, maximum, average, and variance of the optimal values ​​obtained from 500 iterations, and considering the algorithm iteration graphs for each benchmark function, it can be seen that the optimization algorithm of this invention has an optimization effect on the basic Fruit Fly Algorithm, exhibiting faster convergence speed and higher minimum search capability. After calculating the time complexity, the improved GSAFOA and GSAFOA have the same time complexity, without adding unnecessary computational burden, demonstrating the superior performance of the two improved algorithms. With the objective function of minimizing the self-weight of the main girder structure of the bridge erecting machine and the constraints of strength and stiffness design requirements, three algorithms were used for optimization design. The results of the design calculations were then compared. The cross-sectional area of ​​the main girder optimized by the GSAFOA algorithm was reduced by 3.64%, the cross-sectional area optimized by DSEFOA was reduced by 2.33%, and the cross-sectional area optimized by the basic fruit fly algorithm was reduced by 1.33%. The mass of the main girder of the bridge erecting machine was reduced, which can effectively save manufacturing costs.

[0202] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the main beam of a bridge erecting machine based on a golden sine hybrid fruit fly optimization algorithm, characterized in that, Includes the following steps: Step 1: Using the cross-sectional area of ​​the main beam as the optimization objective, determine the objective function and the corresponding key parameters to be optimized; Step 2: Optimize the key parameters using the Golden Sine Hybrid Fruit Fly Optimization Algorithm, specifically including the following steps: Step 2.1: Initialize parameters; Step 2.2: Initialize the search direction and distance of the fruit flies, perform basic fruit fly transformations, and calculate the fitness of each individual fruit fly; Step 2.3: Based on the fitness of each individual fruit fly, determine the best and worst fruit fly individuals and their coordinates in the current fruit fly population; Step 2.4: Based on the coordinates of the best and worst fruit fly individuals, calculate the distance between each fruit fly individual in the population and the best and worst fruit fly individuals, respectively. and ; Step 2.5: If If the fruit fly individuals are assigned to a better subgroup, the positions of the individuals in the better subgroup are updated using the golden sine algorithm; if If the corresponding fruit fly individual is assigned to the inferior subgroup, the position of the fruit fly individual in the inferior subgroup is updated by chaotic mapping. The formula for updating the position of individuals in the superior subgroup using the golden sine algorithm is as follows: ; In the formula, Indicates individual fruit flies In the t Position at the next iteration individual fruit flies i In the t The optimal position in the second generation. For [0, The random number within the interval determines the individual fruit flies in the next iteration. i The distance traveled For [0, A random number between [a certain number] is used to control the individual in the next iteration. i The direction of movement, and The coefficients are derived from the golden ratio t. , ; Step 2.6: Merge the better and worse subgroups, calculate the fitness of each fruit fly individual at its new position, and record the fruit fly individual with the best fitness in the fruit fly population and its position. Step 2.7: Dynamically update the food source location through hybrid mutation, calculated using the following formula: ; in, Let F be the new location of the food source after mixing and mutation, k be the shrinkage factor, which is a random number uniformly distributed between [0,1], and F be the optimal food source location in the current iteration. This represents the Cauchy probabilistic objective function; It is represented as a random number uniformly distributed within the interval [0.1], where M represents the centroid of the neighborhood; Step 2.8: Recalculate the fitness of each fruit fly individual based on the updated food source location; record the optimal fitness value and the coordinates of the corresponding fruit fly individual, and make the fruit fly swarm fly towards that location using visual advantage; Step 2.9: Determine if the maximum number of iterations has been reached. If not, return to step 2.2 and repeat steps 2.2 to 2.8 until the iteration ends, then output the optimal solution. Step 3: Based on the optimization results of the key parameters, perform verification calculations on the strength and stiffness of the main beam structure.

2. The method for optimizing the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm according to claim 1, characterized in that, In step 1, the expression for the objective function is: ; in, The width of the upper flange. The thickness of the upper flange. The web height, The thickness of the lower flange plate. The width of the lower flange. , These are the thicknesses of the two web plates, respectively. The distance between the webs; key parameters to be optimized include ~ .

3. The method for optimizing the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm according to claim 1, characterized in that, In steps 2.2 and 2.5, the specific formulas for performing the basic fruit fly transformation are as follows: in, and Let represent the coordinates of the i-th fruit fly individual after the basic fruit fly transformation. and Let represent the coordinates of the i-th fruit fly individual before the basic fruit fly transformation. RandomValue Indicates the search distance; In step 2.4, the distances between each individual fruit fly in the population and the best and worst fruit fly individuals are... and The calculation formula is: in, and Let represent the distances between the i-th fruit fly individual and the best and worst fruit fly individuals, respectively; , Represents the coordinates of the optimal fruit fly individual. and This represents the coordinates of the worst-performing fruit fly individual.

4. The method for optimizing the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm according to claim 3, characterized in that, In step 2.2, the fitness of each individual fruit fly is calculated as follows: First, calculate the distance between the individual fruit fly and the origin. Then, calculate the individual's taste concentration judgment value using the following formula: in, This represents the distance between the i-th fruit fly individual and the origin. This represents the taste concentration judgment value of the i-th fruit fly individual; Then, the taste concentration determination value Substitute the flavor concentration into the flavor concentration determination function to calculate the flavor concentration of each individual fruit fly as its fitness.

5. The method for optimizing the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm according to claim 1, characterized in that, In step 2.3, the method for determining the best and worst fruit fly individuals in the current fruit fly population is as follows: in, This represents the optimal fruit fly individual. This indicates the worst-quality fruit fly individual. This represents the fitness value of the i-th fruit fly individual.

6. The method for optimizing the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm according to claim 1, characterized in that, Step 2.8 specifically includes the following steps: Determine whether the optimal fitness value obtained after updating the food source location through mixed mutation is less than the optimal fitness value of the fruit fly population obtained before the update. If yes, update the food source location; otherwise, do not update and retain the original food source location information. Then, the optimal fitness value and its corresponding coordinates of the fruit fly individuals are recorded and preserved, and the fruit fly swarm is made to fly to that location using vision.

7. The method for optimizing the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm according to claim 1, characterized in that, In step 2.1, the initialization parameters include the population size. Sizepop Maximum number of iterations Maxgen The location of the fruit fly population and the fitness variance threshold δ.

8. The method for optimizing the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm according to claim 1, characterized in that, In step 2.5, the formula for updating the position of individual fruit flies in the superior subgroup using chaotic mapping is as follows: in, and These represent the control parameters of the i-th dimension variable, , Indicates the location of individual fruit flies , Chaotic variables obtained by performing chaotic mapping and Indicates according to , The intermediate variables obtained , This indicates the new location of an individual fruit fly within the search space.

9. The method for optimizing the main beam of a bridge erecting machine based on the golden sine hybrid fruit fly optimization algorithm according to claim 1, characterized in that, In step 3, the specific method for verifying the strength and stiffness of the main beam structure is as follows: Calculate the deformation, strength, and static stiffness under working conditions one and two, and determine whether they meet the design requirements; working condition one refers to the lifting trolley being located at both ends of the main beam; working condition two refers to the crane being at the mid-span position.

Citation Information

Patent Citations

  • Fruit fly cantilever beam variable measurement method based on logical chaotic mapping and self-adaptive step size

    CN109583020A

  • Ground unmanned platform path planning method based on evolutionary fruit fly optimization algorithm

    CN113495564A