Thrust distribution method of ship dynamic positioning system based on improved genetic algorithm

Abstract

The invention belongs to the technical field of ocean engineering ship dynamic positioning system control, and discloses a ship dynamic positioning system thrust distribution method based on an improved genetic algorithm, which comprises the following steps: establishing a ship dynamic positioning system thrust distribution model; setting parameters, and generating an initial population; calculating the fitness value of each particle; elite individuals are searched; elite individuals are directly copied to a new generation of population; selecting a part of individuals as parent individuals; pairing the parent generations in pairs to generate offspring generations; applying non-uniform variation to the azimuth angle of the filial generation; offspring generated by crossover and variation is added into a new generation of population; calculating fitness values of individuals in a new generation of population, and reserving better individuals; and repeating the steps until the conditions are met, and outputting an optimal solution. According to the thrust distribution optimization method, the characteristics of the thrust distribution optimization problem are combined, the elitist retention mechanism is introduced into the genetic algorithm to be applied to the thrust distribution optimization process, and the defects that a single algorithm is poor in real-time performance, prone to falling into local optimum, low in convergence speed and the like are overcome.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of marine engineering ship power positioning system control, and particularly relates to a ship power positioning system thrust distribution method based on an improved genetic algorithm. BACKGROUND

[0002] The ship power positioning system is one of the core technologies of modern marine engineering ships, drilling platforms, and special ships such as research vessels, which counteracts environmental disturbances such as wind, waves, and currents by real-time adjustment of the thrust and direction of the propeller, so as to keep the ship at a predetermined position and heading in complex sea conditions. As a core part of the power positioning system, thrust distribution needs to meet the demand for coordinated control of multiple propellers while solving the problem of nonlinear optimization under multiple constraints to achieve the goal of minimizing energy consumption, balancing propeller wear, and improving system stability. At present, the main thrust distribution methods include quadratic programming (QP), sequential quadratic programming (SQP), and heuristic algorithms.

[0003] Chinese invention patent: publication number "CN103092077A", named "thrust distribution method of power positioning system", discloses a thrust distribution method of power positioning system, which adopts the following steps: 1: divide the propeller into several groups, and fix the relative angle of the propeller azimuth angle in the group, but not the absolute angle; 2: use the optimization algorithm to obtain a slowly changing propeller azimuth angle from the low-frequency part of the control command, and then use the optimization algorithm to obtain the thrust of the propeller; the specific steps are: input the control command τc,k at time k into the low-pass filter composed of a first-order inertia link, then further discretize to obtain the low-frequency part τk+1; calculate the thrust size f k and the propeller azimuth angle αk of each propeller by using the sequential quadratic programming method from the control command τc,k; then apply the quadratic programming method to obtain the thrust f k of the propeller; and add the thrust reduction coefficient p in the formula of the target function and part of the constraint conditions. This technical solution optimizes the thrust distribution of the power positioning system by using the sequential quadratic programming method. However, this technical solution has the disadvantages of large calculation amount of thrust distribution algorithm, easy to fall into local optimum, poor robustness, etc.

[0004] Chinese invention patent: publication number "CN102385665A", named "ship power positioning system thrust distribution method using genetic algorithm", discloses a ship power positioning system thrust distribution method, which includes: step 31, converting the thrust distribution problem into an optimization problem; step 32, solving the optimization problem of thrust using genetic algorithm; the objective function of the optimization problem is the fitness function in the genetic algorithm, and the constraint condition of the optimization problem is the solution space in the genetic algorithm. This technical solution optimizes the thrust distribution of the power positioning system by using the genetic algorithm, but this technical solution has the problems of slow convergence speed, large solution quality fluctuation, and insufficient real-time performance. SUMMARY

[0005] In order to solve the problems of large amount of calculation, easy to fall into local optimum, poor robustness, slow convergence speed, large solution quality fluctuation and insufficient real-time performance of the existing thrust distribution method, the application aims to provide a thrust distribution method for a ship power positioning system based on an improved genetic algorithm, which introduces an elite reservation strategy into the genetic algorithm, and forces the optimal individual in the current population to be directly entered into the next generation population without being affected by the crossover, mutation or selection operation in the evolution process of each generation of the traditional genetic algorithm. The thrust distribution algorithm can reduce the amount of calculation, speed up the convergence speed and enhance the robustness. Based on considering the motion of three degrees of freedom of the ship, the thrust distribution algorithm can reduce the energy consumption and wear of the propeller and improve the performance of the power positioning system.

[0006] The application is implemented by the following technical solutions:

[0007] Step 1, obtaining a ship power positioning system thrust distribution model according to the ship propeller layout, and the formula is as follows:

[0008] τ=B(α i )·T

[0009] Wherein τ=[F x ,F y ,M z ] T is the instruction thrust vector, F x is the longitudinal resultant force, F y is the transverse resultant force, M z is the yaw moment; T=[T1,T2,...T n ] T is the n-propeller thrust amplitude vector; B(α i ) is a thrust configuration matrix composed of the propeller azimuth angle α i and the position coordinates (l xi , l yi );

[0010] Step 2, setting the genetic algorithm parameters;

[0011] Step 3, randomly generating an initial population in a real number coding mode;

[0012] Step 4, calculating the fitness value of each individual, and the formula of the fitness function is as follows:

[0013] min J=T T WT+s T Qs+(α-α0) T Ω(α-α0)+m|T1-T2|

[0014] Wherein J represents the fitness value of each individual; W is a weight matrix; T is a vector of n thruster thrust magnitude; Q is a weight matrix; s is a thrust synthesis error vector; Ω is a weight matrix; m is a weight coefficient; T1 and T2 are two thrust magnitudes in the same group of thrusters;

[0015] Step 5, compare the fitness values of the individuals, sort the fitness values of all individuals in the population in ascending order, select the top 10% of the individuals as the elite population, and the individual with the smallest fitness value is the current optimal individual. If there are multiple individuals with the same minimum fitness value, take the top 10% of the individuals in order;

[0016] Step 6, directly copy the elite population in the current population obtained in step 5 to the new generation population;

[0017] Step 7, select part of the individuals as parent individuals for generating the next generation;

[0018] Step 8, pair the parent individuals obtained in step 7 two by two, perform binary crossover to generate offspring;

[0019] Step 9, apply non-uniform mutation to the azimuth angle of the offspring, and the mutation amplitude decreases with the number of iterations;

[0020] Step 10, add the offspring generated by the crossover and mutation in steps 8 and 9 to the new generation population, and the positions of the elite individuals in the new generation population remain unchanged;

[0021] Step 11, calculate the fitness value of each individual in the new generation population;

[0022] Step 12, repeat steps 5 to 11 until the fitness value of the optimal individual meets the set threshold or the maximum number of iterations is reached, and output the optimal individual as the thrust magnitude of each thruster and its azimuth angle.

[0023] As a further optimization, the constraint condition of the fitness function in step 4 is as follows:

[0024]

[0025] Wherein τ d represents the expected command thrust vector; B represents the thrust configuration matrix; s represents the thrust synthesis error vector; A represents the linear inequality constraint matrix; b represents a vector; α represents the thruster azimuth angle; α min / α max represents the minimum / maximum value of the thruster azimuth angle; α0 represents the azimuth angle at the previous time; Δα min / Δα max represents the minimum / maximum value of the azimuth angle change rate; T represents the thruster thrust magnitude vector; T min / T maxTmin / max represents the minimum / maximum value of the thrust; T0 represents the thrust vector at the previous time; AT min / AT max Tmin / max represents the minimum / maximum value of the thrust rate; s min / s max Tmin / max represents the minimum / maximum value of the thrust distribution error.

[0026] As a further preferred, the method of selecting parent individuals in step 7 is tournament selection, that is, 3 individuals are randomly selected each time, and the individual with the minimum fitness value is retained as the parent individual.

[0027] As a further preferred, the parent individuals in step 8 are paired two by two, and the formula for generating offspring by binary crossover is as follows:

[0028]

[0029] In the formula, T child is the generated offspring individual; T parent1 and T parent2 are the two parent individuals respectively; and β is a random number subject to uniform distribution.

[0030] As a further preferred, the formula for applying non-uniform mutation to the azimuth of the offspring in step 9 is as follows:

[0031]

[0032] In the formula, α new is the offspring azimuth after mutation; α old is the offspring azimuth before mutation; Δα is the maximum mutation range of the azimuth; t is the current iteration number; and N is the maximum iteration number.

[0033] As a further preferred, the parameters of the genetic algorithm in step 2 include: population size M, maximum iteration number N, crossover probability p c , mutation probability p m , and elite retention ratio k.

[0034] As a further preferred, the maximum mutation range Δα of the azimuth in step 9 is in the range of [-5°, 5°].

[0035] As a further preferred, the population size M in step 2 is 50, and the maximum iteration number N is 200.

[0036] As a further preferred, the crossover probability p c in step 2 is 0.85, the mutation probability p m is 0.1, and the elite retention ratio k is 10%.

[0037] The application also provides a computer readable storage medium for storing a computer program, wherein the computer program executes the improved genetic algorithm-based thrust distribution method of a ship dynamic positioning system.

[0038] Compared with the prior art, the application has the following beneficial effects:

[0039] 1. The application combines the characteristics of the thrust distribution optimization problem, introduces the elite reservation mechanism into the genetic algorithm applied to the thrust distribution optimization process, overcomes the shortcomings of single algorithm such as poor real-time performance, easy to fall into local optimum, slow convergence speed, etc., significantly improves the global convergence ability of the algorithm, and improves the thrust distribution accuracy of the ship dynamic positioning system and reduces the thrust distribution error.

[0040] 2. The application improves the traditional genetic algorithm, considers the power consumption of the propeller in the optimization process, and can more effectively find a low-energy consumption thrust distribution scheme through the elite reservation mechanism, thereby reducing the energy consumption of the ship and making the ship dynamic positioning system more efficient and reliable in actual application.

[0041] 3. The application has small calculation amount, can converge to the optimal solution more quickly, reduces unnecessary calculation steps, and improves the calculation efficiency. Compared with the sequential quadratic programming algorithm, the improved genetic algorithm has obvious advantages in global search ability, processing of complex constraint conditions, parallelism, adaptability, calculation efficiency and robustness. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 The application provides an improved genetic algorithm-based thrust distribution method of a ship dynamic positioning system. DETAILED DESCRIPTION

[0043] The advantages and characteristics of the application will be illustrated and explained by the following non-limiting description of preferred embodiments, which are given by way of example only with reference to the accompanying drawings.

[0044] The application provides an improved genetic algorithm-based thrust distribution method of a ship dynamic positioning system. The improved genetic algorithm refers to the introduction of the elite reservation mechanism into the genetic algorithm. The core is to force to reserve the optimal solution in each generation, avoid the performance degradation caused by the randomness of the traditional genetic algorithm, effectively avoid the loss of excellent genes, and significantly improve the convergence speed and optimization accuracy of the algorithm. The method realizes fast convergence, high stability and low energy consumption of thrust distribution in ship dynamic positioning.

[0045] As shown in Figure 1 The application provides an improved genetic algorithm-based thrust distribution method of a ship dynamic positioning system. The specific steps are as follows:

[0046] Step 1, according to the ship propeller layout, the ship power positioning system thrust distribution model is obtained, the formula is as follows:

[0047] τ=B(α i )·T

[0048] This step establishes a thrust distribution equation according to the ship propeller layout, and constructs a ship power positioning system thrust distribution model, the thrust distribution equation is as follows:

[0049] τ=B(α i )·T

[0050] Where τ=[F x ,F y ,M z ] T is the command thrust vector, F x is the longitudinal resultant force, F y is the transverse resultant force, M z is the yawing moment; T=[T1,T2,...T n ] T is the n propeller thrust amplitude vector; B(α i ) is the thrust configuration matrix composed of propeller azimuth angle α i and position coordinates (l xi , l yi ), its expression is:

[0051]

[0052] Where (l xi , l yi ) is the coordinate of the propeller, which represents the position of the i-th propeller.

[0053] Step 2, set the genetic algorithm parameters, including population size M=50, maximum iteration number N=200, crossover probability p c =0.85, mutation probability p m =0.1 and elite retention ratio k=10%.

[0054] The purpose of this step is to initialize the parameters required by the genetic algorithm in combination with the thrust distribution model. The parameters required by the genetic algorithm include the size of the population M=50, which represents how many initial distribution schemes; the maximum iteration number of the genetic algorithm is set to N=200, which represents the number of searches in the distribution scheme space; the crossover probability p c =0.85 and the mutation probability p m =0.1 represent the exploration and development ability of the algorithm; the elite retention ratio k=10%: represents that the top k individuals of each generation are directly entered into the next generation;

[0055] Step 3: Use real number encoding to randomly generate the initial population.

[0056] Encoding is a method of representing an individual's location. Real-number encoding means each gene represents a variable, and each thruster's thrust amplitude T... i and azimuth α i As the variables to be optimized, they constitute the continuous parameter space of an individual, with each individual represented as a D-dimensional vector (D = 2n, where n is the number of thrusters), and the first n dimensions representing the thrust amplitude T. i The last n dimensions are the azimuth angle α. i Randomly generate the initial population, satisfying the following constraints:

[0057] T min ≤T i ≤T max ,α min ≤α i ≤α max

[0058] Step 4: Calculate the fitness value for each individual. The formula for the fitness function is as follows:

[0059] minJ = T T WT+s T Qs+(α-α0) T Ω(α-α0)+m|T1-T2|

[0060] In this invention, the fitness function of each individual is the objective function for thrust allocation. The objective function primarily aims at optimizing thruster energy consumption, while also considering reducing the deviation between the desired commands of the control system and the ship's output commands, minimizing thruster wear, and achieving heading optimization. The general form of the objective function is:

[0061] minJ = T T WT+s T Qs+(α-α0) T Ω(α-α0)+m|T1-T2|

[0062] Where J represents the fitness value of each individual, the first term in the objective function represents the total power consumption of each thruster, W is the weight matrix, and T is the thrust amplitude vector of n thrusters; the second term is the error penalty term, Q is the weight matrix, the magnitude of which determines which direction is optimal in the three degrees of freedom, and s is the thrust synthesis error vector; the third term is to prevent severe rudder angle vibration, reduce thruster wear, and thus extend thruster life, Ω is the weight matrix of the error term; the last term avoids excessive difference in thrust between the two bow sides to prevent thrust from canceling each other out and causing energy waste, m is the weight coefficient; T1 and T2 are the two thrust amplitudes in the same group of thrusters;

[0063] The constraints of the fitness function are as follows:

[0064]

[0065] where τ d denotes the desired command thrust vector; B denotes the thrust configuration matrix; s denotes the thrust synthesis error vector for handling infeasibility problems when thrust saturation or failure occurs; A denotes the linear inequality constraint matrix; b denotes a vector; a denotes the propeller azimuth angle (usually 0° with respect to the ship longitudinal axis, 0° to 360° clockwise); a min / a max denotes the minimum / maximum value of the propeller azimuth angle, which is determined by the propeller type: a min = 0°, a max = 360° for a Z-peller; a min = -30°, a max = 30° for a tunnel thruster; a0denotes the azimuth angle vector at the last time instant; a min / a max denotes the minimum / maximum value of the azimuth angle rate of change; T denotes the propeller thrust magnitude vector; T min / T max denotes the minimum / maximum value of the thrust; T0denotes the thrust vector at the last time instant; T min / T max denotes the minimum / maximum value of the thrust rate of change; s min / s max denotes the minimum / maximum value of the thrust allocation error.

[0066] The consideration items of the constraint conditions of the fitness function of the individual of the present application include: thrust magnitude, thrust rate of change, azimuth angle rate of change, thrust region.

[0067] where the thrust synthesis equation constraint is as follows:

[0068] T d = BT + s

[0069] The thrust synthesis equation constraint can describe the dynamic balance relationship between the propeller thrust synthesis and the ship demand thrust, T d denotes the desired command thrust vector, B denotes the thrust configuration matrix, and s denotes the thrust synthesis error vector for handling infeasibility problems when thrust saturation or failure occurs;

[0070] The linear inequality constraint is as follows:

[0071] AT≤b

[0072] The linear inequality constraint globally limits the linear combination of thrust allocation, A denotes the linear inequality constraint matrix, and b denotes a vector;

[0073] The azimuth angle range constraint is as follows:

[0074] α min ≤α≤α max

[0075] Azimuth angle range constraint limits the mechanically adjustable range of the propeller azimuth angle, α is the propeller azimuth angle (usually with the ship longitudinal axis as 0°, clockwise 0° to 360°); α min / α max Determined by the propeller type, if it is a full-revolving propeller, then α min = 0°, α max = 360°; if it is a tunnel propeller, then α min = -30°, α max = 30°;

[0076] Azimuth angle change rate constraint:

[0077] Δα min ≤α-α0≤Δα max

[0078] Azimuth angle change rate constraint limits the maximum adjustment amplitude of the azimuth angle within the adjacent control period, α0 is the azimuth angle at the last time; Δα min / Δα max represents the minimum / maximum value of the azimuth angle change rate;

[0079] Thrust amplitude constraint:

[0080] T min ≤T≤T max

[0081] Thrust amplitude constraint limits the physically feasible range of the propeller thrust, T min / T max respectively represent the minimum controllable thrust and the rated maximum thrust;

[0082] Thrust change rate constraint:

[0083] ΔT min ≤T-T0≤ΔT max

[0084] Thrust change rate constraint limits the thrust adjustment speed within the adjacent control period, T0 is the thrust at the last time; ΔT min / ΔT max Determined by the motor response characteristics; for a dynamic model, if the control period is Δt, the constraint can be converted to:

[0085]

[0086] Slack variable constraint:

[0087] s min≤ s ≤ s max

[0088] Relaxing variable constraint limits the allowable deviation range of thrust synthesis error, s min / s max Indicates the minimum / maximum value of thrust distribution error, limits the allowable range of thrust distribution error, when the thruster fails, allows sacrificing the secondary degree of freedom (such as sway) to prioritize the key degree of freedom (such as yaw).

[0089] Step 5, compare the fitness values of individuals, sort all individual fitness values in the population in ascending order, select the top 10% of individuals as the elite population, and the individual with the smallest fitness value is the current optimal individual. If there are multiple individuals with the same minimum fitness value, take the top 10% of individuals in order;

[0090] Step 6, directly copy the elite population in the current population obtained in step 5 to the new generation population;

[0091] Step 7, select part of the individuals as parent individuals for generating the next generation, preferably using tournament selection method, randomly select 3 individuals each time, and keep the individual with the smallest fitness value as the parent individual for generating the next generation.

[0092] This step is a selection operation, according to the fitness value of the individual, select part of the individuals as the parent for generating the next generation. Using tournament selection method, randomly select 3 individuals each time, and keep the individual with the smallest fitness value. The purpose of selection is to give the individual with small fitness value a greater chance of being selected, so as to pass on its excellent genes.

[0093] Step 8, pair the parent individuals obtained in step 7 two by two, perform binary crossover (SBX), generate offspring, the formula is as follows:

[0094]

[0095] In the formula, T child is the generated offspring individual; T parent1 and T parent2 are two parent individuals respectively; β is a random number subject to uniform distribution;

[0096] This step is a crossover operation, which pairs the selected parent individuals two by two to generate new offspring through crossover operation.

[0097] Step 9, apply non-uniform mutation to the azimuth angle of the offspring, and the mutation amplitude decreases with the number of iterations, the formula is as follows:

[0098]

[0099] Where, α newis the azimuth angle of the offspring after mutation; a is the maximum variation range of the azimuth angle, and takes a random value in [-5°, 5°]; t is the current iteration number; and N is the maximum iteration number. old is the azimuth angle of the offspring before mutation, a is the maximum variation range of the azimuth angle, and takes a random value in [-5°, 5°]; t is the current iteration number; and N is the maximum iteration number.

[0100] Step 10: The offspring generated by the crossover and mutation in steps 8 and 9 are added to the new generation population, and the positions of the elite individuals in the new generation population remain unchanged.

[0101] Step 11: The fitness values of each individual in the new generation population (including the 10% elite individuals copied in step 6 and the offspring individuals generated by the crossover and mutation in step 10) are calculated, in preparation for the next round of sorting and elite selection.

[0102] Step 12: Steps 5 to 11 are repeated until the fitness value of the optimal individual meets the set threshold or the maximum iteration number is reached, and the optimal individual is output as the thrust amplitude and azimuth angle of each propeller.

[0103] Those skilled in the art can understand that all or part of the steps of the above method embodiments can be completed by program instructions and related hardware, the aforementioned program instructions can be stored in a computer-readable storage medium, and the program instructions execute the steps including the above method embodiments when executed; and the aforementioned storage medium includes a mobile storage device, a read-only memory (ROM), a magnetic disc or an optical disc, and various media that can store program codes.

[0104] In addition to the above embodiments, the present application can also have other implementation manners, and any technical solutions formed by equivalent replacement or equivalent transformation fall within the protection scope of the present application.

Claims

1. A method for thrust allocation in a ship dynamic positioning system based on an improved genetic algorithm, characterized in that: Comprising the following steps: Step 1, obtaining a ship power positioning system thrust distribution model according to the ship propeller layout, the formula is as follows: τ = B(α i ) · T where τ = [F x , F y , M z ] T is the commanded thrust vector, F x is the longitudinal resultant force, F y is the lateral resultant force, M z is the yawing moment; T = [T1, T2,... T n ] T is the n thruster thrust magnitude vector; B(a i ) is the thrust configuration matrix composed of the thruster azimuth angle a i and the position coordinates (l xi , l yi ). Step 2, setting genetic algorithm parameters; Step 3, randomly generating an initial population in a real number coding mode; Step 4, calculating the fitness value of each individual, and the formula of the fitness function is as follows: minJ = T T WT + s T Qs + (a - a0) T Ω(a - a0) + m|T1 - T2| Wherein J represents the fitness value of each individual; W is a weight matrix; T is an n-propeller thrust amplitude vector; Q is a weight matrix; s is a thrust synthesis error vector; Ω is a weight matrix; m is a weight coefficient; T1 and T2 are two thrust amplitudes in the same group of propellers; Step 5, comparing the fitness values of individuals, sorting all individual fitness values in the population in ascending order, and selecting the top 10% of individuals as the elite population, wherein the individual with the smallest fitness value is the current optimal individual. If there are multiple individuals with the same minimum fitness value, the first 10% of individuals are selected in order; Step 6, copying the elite population in the current population obtained in step 5 directly into the next generation population; Step 7, selecting part of the individuals as parent individuals for generating the next generation; Step 8, pairing the parent individuals obtained in step 7 two by two, performing binary crossover to generate offspring; Step 9, applying non-uniform mutation to the azimuth of the offspring, and the mutation amplitude decreases with the number of iterations; Step 10, adding the offspring generated by step 8 and step 9 to the new generation population, and the positions of the elite individuals in the new generation population remain unchanged; Step 11, calculating the fitness value of each individual in the new generation population to prepare for the next round of sorting and elite selection; Step 12, repeating steps 5 to 11 until the fitness value of the optimal individual meets the set threshold or the maximum number of iterations is reached, and outputting the optimal individual as the thrust amplitude of each propeller and its azimuth.

2. The method for thrust allocation of a ship's dynamic positioning system based on improved genetic algorithm according to claim 1, characterized in that: The constraint condition of the fitness function in step 4 is as follows: where τ d denotes the desired command thrust vector; B denotes the thrust configuration matrix; s denotes the thrust allocation error vector; A denotes the linear inequality constraint matrix; b denotes a vector; a denotes the thruster azimuth angle; a min / a max denotes the minimum / maximum value of the thruster azimuth angle; a0 denotes the azimuth angle at the previous time instant; a min / a max denotes the minimum / maximum value of the azimuth angle rate; T denotes the thruster thrust magnitude vector; T min / T max denotes the minimum / maximum value of the thrust; T0 denotes the thrust vector at the previous time instant; T min / T max denotes the minimum / maximum value of the thrust rate; s min / s max denotes the minimum / maximum value of the thrust allocation error.

3. The method for thrust allocation of a ship's dynamic positioning system based on improved genetic algorithm according to claim 2, characterized in that: The method for selecting parent individuals in step 7 is tournament selection, that is, randomly selecting 3 individuals each time, and retaining the individual with the smallest fitness value as the parent individual.

4. The method for thrust allocation of a ship's dynamic positioning system based on improved genetic algorithm according to claim 3, characterized in that: The formula for pairing the parent individuals two by two in step 8 to generate offspring is as follows: where T child is the generated offspring individual; T parent1 and T parent2 are the two parent individuals, respectively; and β is a random number following a uniform distribution.

5. The method for thrust allocation of a ship's dynamic positioning system based on improved genetic algorithm according to claim 4, characterized in that: The formula for applying non-uniform mutation to the azimuth of the offspring in step 9 is as follows: wherein α new is the azimuth angle of the offspring after mutation; α old is the azimuth angle of the offspring before mutation; Δα is the maximum variation range of the azimuth angle; t is the current iteration number; and N is the maximum iteration number.

6. The method of thrust allocation for a ship's dynamic positioning system based on an improved genetic algorithm according to any one of claims 1 to 5, characterized in that: The parameters of the genetic algorithm in step 2 include: population size M, maximum iteration number N, crossover probability p c , mutation probability p m , and elite reservation ratio k.

7. The method for thrust allocation of a ship's dynamic positioning system based on improved genetic algorithm according to claim 5, characterized in that: The maximum variation amplitude Δα of the azimuth in step 9 is in the range of [-5°, 5°].

8. The method for thrust allocation of a ship's dynamic positioning system based on improved genetic algorithm according to claim 6, characterized in that: The population size M in step 2 is 50, and the maximum number of iterations N is 200.

9. The method for thrust allocation of a ship's dynamic positioning system based on improved genetic algorithm according to claim 6, characterized in that: The crossover probability p in step 2 c is 0.85, the mutation probability p is 0.1, and the elite reservation ratio k is 10%. m is 0.85, the mutation probability p is 0.1, and the elite reservation ratio k is 10%.

10. A computer-readable storage medium, characterized in that: The computer readable storage medium is used to store a computer program, and the computer program executes the ship power positioning system thrust distribution method based on the improved genetic algorithm of claim 7.

Images