Ship ballast water adjusting method based on improved tornado optimization algorithm

Abstract

The invention discloses a ship ballast water adjusting method based on an improved tornado optimization algorithm, and the method comprises the steps: S1, obtaining ship ballast tank information and a current ballast water state, and constructing a target function according to the ship ballast tank information and the current ballast water state; s2, constructing a ballast water constraint condition according to the target function; s3, designing an improved tornado optimization algorithm and solving the target function in combination with constraint conditions of ballast water to obtain an optimal solution as a current ballast water adjustment scheme; and S4, adjusting the ballast water according to the ballast water adjusting scheme. Through the improved tornado optimization algorithm, the calculation efficiency and optimization precision of ballast water adjustment are remarkably improved, and the ship navigation safety and economy are effectively guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of ballast water scheduling, and in particular to a ship ballast water regulation method based on an improved tornado optimization algorithm. Background Art

[0002] Ballast water allocation plays a vital role in a ship's navigation. Ballast water is used to adjust the ship's center of gravity, draft, and stability, ensuring it remains stable and safe under varying loads and sea conditions. Traditional ballast water allocation methods typically rely on manual calculations or pre-set rules. Crew members manually adjust the water volume in each ballast tank based on the ship's real-time status to control the ship's draft and stability. However, traditional manual ballast water allocation methods have significant drawbacks such as low efficiency, high computational burden, and insufficient accuracy when addressing complex operational requirements.

[0003] To overcome the shortcomings of traditional ballast water allocation methods, ballast water allocation technology has gradually developed towards automation and intelligence in recent years. For example, some ships are equipped with automatic ballast water allocation systems, which use computers and control systems to precisely control the water volume in each ballast tank. These improvements have improved the efficiency and accuracy of ballast water allocation to a certain extent, but the following limitations still exist: First, some automatic ballast water allocation systems rely on traditional optimization methods such as graph theory and moment balance to distribute water. Although effective, these methods are computationally complex, especially in multi-compartment or multi-objective optimization scenarios. The solution speed is slow and it is difficult to respond to the dynamic changes of the ship in real time. Second, many optimization methods are prone to falling into local optimal solutions during the solution process, resulting in the allocation solution failing to reach the global optimal solution. Consequently, the allocation of ballast water cannot be minimized, resulting in water waste and affecting the ship's stability and fuel efficiency. In addition, existing ballast water allocation methods often rely on preset algorithms and rules and lack the ability to adapt to the ship's real-time operating status and external environment. This makes the allocation process inflexible and unable to cope with complex and changing sea conditions and operating conditions. These limitations mean that the adaptability and computational efficiency of existing automatic ballast water allocation systems under complex conditions still have a lot of room for improvement. Therefore, there is a need for a method that can efficiently and accurately perform optimal ballast water allocation under complex sea conditions. Summary of the Invention

[0004] The present invention provides a ship ballast water regulation method based on an improved tornado optimization algorithm, so as to overcome the problems of low adaptability and computational efficiency of the existing ballast water allocation system and easy falling into the local optimal solution.

[0005] In order to achieve the above object, the technical solution of the present invention is:

[0006] A ship ballast water regulation method based on an improved tornado optimization algorithm, comprising:

[0007] S1. Acquire ship ballast tank information and current ballast water status, and construct an objective function based on the ship ballast tank information and the current ballast water status;

[0008] S2. Construct ballast water constraints based on the objective function;

[0009] S3. Design an improved Tornado optimization algorithm and solve the objective function in combination with the ballast water constraints to obtain the optimal solution as the current ballast water regulation plan;

[0010] S4. Regulate the ballast water according to the ballast water regulation plan.

[0011] Furthermore, the expression of the objective function is constructed as follows:

[0012]

[0013] Where, P is the ballast water transfer volume; Δh ji is the allocation quantity of ballast tank i; S i is the bottom area of ​​ballast tank i; ρ sw is the density of seawater; n is the number of ballast tanks;

[0014] Wherein, the allocation amount Δh of the ballast compartment i is calculated ji , whose expression is:

[0015] Δh ji =h ji -h (j-1)i (2)

[0016] Where h ji is the water level of ballast tank i at stage j; h (j-1)i is the water level of ballast tank i at stage j-1.

[0017] Furthermore, the ballast water constraint conditions include: hull balance constraint conditions, ballast tank water level constraint conditions, ballast water initial loading capacity constraint conditions, and ballast total water capacity detection constraint conditions.

[0018] Furthermore, the ballast water constraint conditions are specifically as follows:

[0019] S21. The hull balance constraint condition refers to the hull maintaining moment balance under the combined action of multiple moments. Its expression is:

[0020] M B +M H +M T +M R =0(3)

[0021] Where M Bis the ship ballast moment; M T is the weight's pitch moment; M H is the heeling moment of the weight; M R is the restoring torque;

[0022] The hull balance also satisfies the hull's longitudinal and transverse angles θ≤[θ], Among them, [θ] and are the permissible thresholds for pitch and heel angles, respectively;

[0023] S22. The ballast tank water level constraint condition means that the ballast tank water level cannot be higher than the cabin height and satisfies 0≤h j ≤H j ; Among them, h j is the ballast tank water level; H j The maximum water level allowed in the ballast tank;

[0024] S23. The initial ballast water loading constraint condition means that the initial ballast water loading quantity satisfies Q min ≤Q i ≤Q max Among them, Q min is the minimum ballast water loading; Q max is the maximum ballast water loading capacity;

[0025] S24. The constraint condition for the detection of total ballast water volume is that the total ballast water volume must remain unchanged before and after allocation.

[0026] Furthermore, the specific process of solving the objective function by combining the improved tornado optimization algorithm with the ballast water constraint condition includes:

[0027] S31. Initialize the optimization population by introducing Cat chaos mapping and reverse learning strategy; the individuals in the initialized optimization population are the water level adjustment schemes for the ballast tanks; wherein the initialization process is:

[0028] S311, generating a chaotic sequence according to the Cat chaotic map and mapping it to the upper and lower bounds of the variable [l j ,u j ] to obtain the initial population, the upper and lower bounds of the variables are the upper and lower bound solutions of the ballast tank capacity;

[0029] S312: Calculate the reverse solution of the initial population solution according to the reverse learning strategy, and its expression is:

[0030] X' i =L i +U i -X i (4)

[0031] Where, L i 、Ui are the upper and lower bounds; X i is the initial population solution; X' i is the reverse population solution;

[0032] S313, merging the initial solution and the reverse population solution into a mixed solution set {X' i ,X i};

[0033] S32. Calculate the fitness of all solutions in the mixed solution set according to the objective function and the ballast water constraint condition, and the expression is:

[0034]

[0035] Where P is the ballast water transfer amount obtained by the objective function; λ is the penalty function, and the expression defining the penalty function is:

[0036] λ=λ0·(1+cos(πt / T) (6)

[0037] Where: λ0 is the initial coefficient of the penalty function; t is the number of iterations;

[0038] The mixed solution set {X i ',X i Sort by fitness from small to large, and select the top N candidate solutions with the smallest fitness as the final initial population;

[0039] According to the fitness, the final initial population is divided into tornado-level candidate solutions, thunderstorm-level candidate solutions and storm-level candidate solutions; wherein the tornado-level candidate solutions are the first n candidates with the smallest fitness. o solutions, the thunderstorm-level candidate solution is n solutions whose fitness is greater than that of the tornado-level candidate solution. t solutions, the storm-level candidate solution is n of the remaining fitness w a solution;

[0040] S33. Dynamically update the speed of the storm level candidate solution by introducing a nonlinear inertia weight. The speed is the deployment direction that determines the deployment amount of the ballast tank. The update expression is:

[0041]

[0042] Where i = 1, 2, ..., n w Indicates that the population size is n w Storm Index; is the new velocity vector of the candidate solution for the i-th storm level; is the current velocity vector of the candidate solution for the i-th storm level; rand is a random number uniformly distributed in the interval [0,1]; η is the convergence factor; ω1 is the nonlinear inertia weight; μ represents the fuzzy adaptive random number; R l and R r are the curvature radii of the storm tracks in the Northern Hemisphere and the Southern Hemisphere respectively; c is a random number in different ranges; define f, CF l ,CF r The expression is as follows:

[0043] f=2·Ω×sin(-1+2·rand) (8)

[0044] Where Ω represents the angular velocity; -1+2·rand represents the random value of the specified dimension, which is the index of the water level adjustment amount of a single cabin;

[0045]

[0046]

[0047] Where, is the component of the pressure gradient force in the current direction of the i-th storm at the t-th iteration;

[0048] S34. Adopting an adaptive external disturbance factor, reconstructing the evolution rule of the storm-level candidate solution to the tornado-level candidate solution; the evolution rule is:

[0049] S341. Update the position of the storm-level candidate solution based on the speed of the storm-level candidate solution. The expression is:

[0050]

[0051] Where, is the next position vector of the candidate solution of storm level i at the t+1th iteration; is the current position vector of the i-th storm level candidate solution at the t-th iteration; is the current position vector of the i-th tornado-level candidate solution at the t-th iteration; The difference between the evolution of the storm-level candidate solution into the tornado-level candidate solution and the random formation of the storm-level candidate solution; rand ω is the index vector of randomly selected storms;

[0052] S342. When the fitness of the storm-level candidate solution after the update is less than the fitness of the dominant thunderstorm-level candidate solution or tornado-level candidate solution with the lowest fitness, the storm-level candidate solution exchanges positions with the dominant thunderstorm-level candidate solution or tornado-level candidate solution to become the dominant candidate solution, and a new storm-level candidate solution is randomly generated.

[0053] The new storm-level candidate solution is independently subjected to boundary verification and fitness calculation, and is added to the population again from the initial stage of step S342;

[0054] S343. If the position is not exchanged, the jump operation is performed, and the expression is:

[0055]

[0056] Where l and u represent the lower and upper limits of the search region, respectively; δ2 represents the sign change; ||·|| represents the norm operator; ν is the defined exponential function; a y represents the exponential parameter;

[0057] S35. Set a tolerance level r and a tolerance threshold R according to the tolerance mechanism, and determine whether the algorithm falls into a local optimal state by using the tolerance level and the tolerance threshold; calculate the tolerance level r, which is expressed as:

[0058]

[0059] Where r is the tolerance level; is the position of the optimal solution of the previous iteration; is the position of the optimal solution of the current iteration;

[0060] When the tolerance level r is greater than or equal to the tolerance threshold R, it is determined that a local optimal solution has been reached; the random storm-level candidate solution is initialized according to differential evolution, and the tolerance level r is reset to zero, and the iterative execution returns to S33; otherwise, the termination judgment is entered;

[0061] S36. When the maximum number of iterations reaches a preset threshold or meets the convergence condition, the optimal solution, i.e., the optimal ballast water allocation plan, is output; otherwise, the process returns to step S33 for iterative execution; the triggering rule for the convergence condition is that the tolerance level returns to zero and becomes equal to the tolerance threshold again.

[0062] The present invention has the following beneficial effects: the present invention optimizes and calculates the ship's ballast water through an improved tornado optimization algorithm, effectively avoiding the problem that traditional methods are prone to falling into local optimality, and can still efficiently search for the global optimal solution under complex multi-constraint conditions; by dynamically adjusting the algorithm convergence strategy, the number of iterations is reduced, and the timeliness requirements of real-time ship control are met while improving the calculation efficiency; multiple safety constraints are integrated into the objective function to ensure the safety indicators of the ship after ballast water adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0064] Figure 1 This is a flow chart of the adjustment method of the present invention;

[0065] Figure 2 This is a flowchart of the tornado algorithm execution of the present invention. DETAILED DESCRIPTION

[0066] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0067] This embodiment provides a ship ballast water regulation method based on an improved tornado optimization algorithm, such as Figure 1 Shown, including:

[0068] S1. Acquire ship ballast tank information and current ballast water status, and construct an objective function based on the ship ballast tank information and the current ballast water status;

[0069] S2. Construct ballast water constraints based on the objective function;

[0070] S3. Design an improved Tornado optimization algorithm and solve the objective function in combination with the ballast water constraints to obtain the optimal solution as the current ballast water regulation plan;

[0071] S4. Regulate the ballast water according to the ballast water regulation plan.

[0072] Specifically, first, the ship's ballast tank information (number of ballast tanks, maximum capacity of each ballast tank, bottom area of ​​the ballast tank, seawater density) and the current ballast water status (initial water volume in the ballast tank, water level height) are obtained to construct a multi-objective optimization function; secondly, based on the ship's stability specifications and mechanical constraints (such as pump flow upper limit, tank capacity limit), physical constraints for ballast water regulation are constructed to ensure the safety and feasibility of the scheme; thirdly, an improved tornado optimization algorithm is designed, which enhances the algorithm's global search capability and convergence speed by introducing Cat chaos mapping and reverse learning strategy, nonlinear inertia weight, external disturbance factor and tolerance mechanism, and efficiently solves the objective function in combination with the constraints to generate the optimal ballast water allocation scheme; finally, the ballast pump is driven to execute flow distribution to achieve rapid and stable floating state of the ship.

[0073] In a specific embodiment, the expression for constructing the objective function is:

[0074]

[0075] Where, P is the ballast water transfer volume; Δh ji is the allocation quantity of ballast tank i; S i is the bottom area of ​​ballast tank i; ρ sw is the density of seawater; n is the number of ballast tanks;

[0076] Wherein, the allocation amount Δh of the ballast compartment i is calculated ji , whose expression is:

[0077] Δh ji =h ji -h (j-1)i (2)

[0078] Where h ji is the water level of ballast tank i at stage j; h (j-1)i is the water level of ballast tank i at stage j-1.

[0079] In a specific embodiment, the ballast water constraint conditions include: hull balance constraint conditions, ballast tank water level constraint conditions, ballast water initial loading capacity constraint conditions, and total ballast water quantity detection constraint conditions.

[0080] The ballast water constraints are specifically:

[0081] S21. The hull balance constraint condition refers to the hull maintaining moment balance under the combined action of multiple moments. Its expression is:

[0082] M B +M H +M T +M R =0 (3)

[0083] Where M B is the ship ballast moment; M T is the weight's pitch moment; M H is the heeling moment of the weight; M R is the restoring torque;

[0084] The hull balance also satisfies the hull's longitudinal and transverse angles θ≤[θ], Among them, [θ] and are the permissible thresholds for pitch and heel angles, respectively;

[0085] S22. The ballast tank water level constraint condition means that the ballast tank water level cannot be higher than the cabin height and satisfies 0≤h j ≤H j ; Among them, h j is the ballast tank water level; H j The maximum water level allowed in the ballast tank;

[0086] S23. The initial ballast water loading constraint condition means that the initial ballast water loading quantity satisfies Q min ≤Q i ≤Q max Among them, Q min is the minimum ballast water loading; Q max is the maximum ballast water loading capacity;

[0087] S24. The constraint condition for the detection of total ballast water volume is that the total ballast water volume must remain unchanged before and after allocation.

[0088] In a specific embodiment, the specific process of solving the objective function by combining the improved tornado optimization algorithm with the ballast water constraint condition includes:

[0089] S31. In the Tornado optimization algorithm, the random initialization of the population makes the initial population unevenly distributed in the solution space and the population diversity is poor. Therefore, in order to maintain population diversity and uniformly position the individuals in the initial population, the optimized population is initialized in the search space of the ballast water tank water level adjustment parameter by introducing Cat chaos mapping and reverse learning strategy. The individuals in the initialized optimized population are the ballast tank water level adjustment scheme. The initialization process is as follows:

[0090] S311, generating a chaotic sequence according to the Cat chaotic map and mapping it to the upper and lower bounds of the variable [l j ,u j ] to obtain the initial population, the upper and lower bounds of the variables are the upper and lower bound solutions of the ballast tank capacity;

[0091] S312, after using Cat chaotic mapping to generate an initial population, use a reverse learning strategy to find a reverse solution for the initial population, thereby improving the quality of the optimal point; calculate the reverse solution of the initial population solution, and its expression is:

[0092] X' i =L i +U i -X i (4)

[0093] Where, L i 、U i are the upper and lower bounds; X i is the initial population solution; X' i is the reverse population solution;

[0094] S313, merging the initial solution and the reverse population solution into a mixed solution set {X' i ,X i};

[0095] S32. Calculate the fitness of all solutions in the mixed solution set according to the objective function and the ballast water constraint condition, and the expression is:

[0096]

[0097] Where P is the ballast water transfer amount obtained by the objective function; λ is the penalty function, and the expression defining the penalty function is:

[0098] λ=λ0·(1+cos(πt / T) (6)

[0099] Where: λ0 is the initial coefficient of the penalty function; t is the number of iterations;

[0100] The mixed solution set {X' i ,X i Sort by fitness from small to large, and select the top N candidate solutions with the smallest fitness as the final initial population;

[0101] According to the fitness, the final initial population is divided into tornado-level candidate solutions, thunderstorm-level candidate solutions and storm-level candidate solutions; wherein the tornado-level candidate solutions are the first n candidates with the smallest fitness. o solutions, the thunderstorm-level candidate solution is n solutions whose fitness is greater than that of the tornado-level candidate solution. t solutions, the storm-level candidate solution is n of the remaining fitness w a solution;

[0102] The difference in fitness is used to determine how many storms each tornado-level candidate solution and thunderstorm-level candidate solution absorbs. The expression is:

[0103]

[0104] Where, f k = is the difference in fitness, where n to is the sum of the number of tornado-level candidate solutions and thunderstorm-level candidate solutions n to =n t +n o ; is the rounding function, is the number of storms that evolve or are designated as specific thunderstorms or tornadoes;

[0105] S33. Dynamically update the speed of the storm-level candidate solution by introducing a nonlinear inertia weight. The speed determines the allocation direction of the ballast tank allocation amount. This update strategy maintains a high weight value in the early stage of the algorithm search to enhance the global exploration capability (improve the efficiency of wind speed traversal search space), and reduces the weight value in the later stage to enhance the local development accuracy (suppress wind speed fluctuations). The specific update rules are:

[0106]

[0107] Where i = 1, 2, ..., n w Indicates that the population size is n w Storm Index; is the new velocity vector of the candidate solution for the i-th storm level; is the current velocity vector of the candidate solution for the i-th storm level; rand is a random number uniformly distributed in the interval [0,1]; η is the convergence factor; ω1 is the nonlinear inertia weight; μ represents the fuzzy adaptive random number; R l and R r are the curvature radii of the storm tracks in the Northern Hemisphere and the Southern Hemisphere respectively; c is a random number in different ranges; define η; ω1; μ; R l 、R r ,c,f,CF l ,CF r The expression is as follows:

[0108] The expression defining the shrinkage factor is:

[0109]

[0110] Where χ represents the acceleration rate of the storm, and after investigation, it was found that χ = ​​4.10; the contraction factor η = 0.7298;

[0111] The expression for defining the nonlinear inertia weight is:

[0112]

[0113] Where t is the number of iterations; T maxis the maximum number of iterations; ω1 is the nonlinear inertia weight, which dynamically coordinates global exploration and local development behaviors to avoid premature convergence and improve convergence accuracy;

[0114] The expression for defining fuzzy adaptive random numbers is:

[0115]

[0116] Where, the expected value of μ is 0.75;

[0117] The expressions defining the radius of curvature of storm tracks in the Northern and Southern Hemispheres are:

[0118]

[0119] The expressions for defining random numbers in different ranges are:

[0120] c=b r ×δ1×ω r (14)

[0121] Where b r is a constant equal to 100,000; δ1 represents a sign change; ω r are random values ​​generated in different ranges; where δ1 and ω are defined r The expression is:

[0122]

[0123] Where, f d Represents a function with values ​​1 and -1; ω min and ω max They are fixed values ​​1 and 4 respectively;

[0124] The expression defining f= is:

[0125] f=2·Ω×sin(-1+2·rand) (17)

[0126] Where Ω represents the angular velocity, and the value of Ω is 0.7292115E-04 radians per second; -1+2 rand represents the random value of the specified dimension, and the specified dimension is the water level adjustment index of a single cabin;

[0127] Defining CF l and CF r The expression is:

[0128]

[0129]

[0130] Where, is the component of the pressure gradient force in the current direction of the i-th storm in the t-th iteration, and its expression is:

[0131]

[0132] Where, is the current position vector of the tornado-level candidate solution at random index ζ in iteration t; is the current position vector of the storm-level candidate solution in the t-th iteration; where ζ is the random index of the tornado-level candidate solution, which is expressed as:

[0133]

[0134] In addition, the following formula is used to constrain Formula 8:

[0135]

[0136] S34. Adopting an adaptive external disturbance factor, reconstructing the evolution rule of the storm-level candidate solution to the tornado-level candidate solution; the evolution rule is:

[0137] S341. Update the position of the storm-level candidate solution based on the speed of the storm-level candidate solution. The expression is:

[0138]

[0139] Where, is the next position vector of the candidate solution of storm level i at the t+1th iteration; is the current position vector of the i-th storm level candidate solution at the t-th iteration; is the current position vector of the i-th tornado-level candidate solution at the t-th iteration; The difference between the evolution of the storm-level candidate solution into the tornado-level candidate solution and the random formation of the storm-level candidate solution; rand ω is the index vector of randomly selected storms;

[0140] Among them, the expression for defining the index vector of the randomly selected storm is:

[0141]

[0142] The expression for defining α is:

[0143] α=|2a y ·rand-rand| (27)

[0144] Where a y Represents the exponential parameter, and its expression is:

[0145]

[0146] Where a0 represents the constant 2;

[0147] The expression for defining ω2 is:

[0148]

[0149] Where ω2 is the Cauchy-Gaussian perturbation, which is used to escape from the local optimum;

[0150] S342, the storm-level candidate solution evolves to the thunderstorm-level candidate solution, and the storm-level candidate solution is assigned to a thunderstorm-level candidate solution, and its expression is:

[0151]

[0152] Where, represents the next position vector of the storm-level candidate solution that evolves into the thunderstorm-level candidate solution at the t+1th iteration; represents the current position vector of the storm-level candidate solution that evolves into the thunderstorm-level candidate solution at the t-th iteration; represents the current position vector of the i-th thunderstorm level candidate solution at the t-th iteration;

[0153] S343: The thunderstorm-level candidate solution evolves toward the tornado-level candidate solution. The thunderstorm-level candidate solution is affected by the tornado-level candidate solution and other thunderstorm-level candidate solutions. The position of the thunderstorm-level candidate solution moves closer to the more optimal solution. The expression is:

[0154]

[0155] Where, represents the next position vector of the thunderstorm-level candidate solution that evolves into the tornado-level candidate solution at the t+1th iteration; They represent the current position vector of the thunderstorm-level candidate solution that evolves into the tornado-level candidate solution at the t-th iteration; represents the position vector of the candidate solution of tornado magnitude at random index ζ; Represents the position vector of the candidate thunderstorm solution at the random index vector p; p is the index vector of the randomly selected candidate thunderstorm solution

[0156] S344. When the fitness of the storm-level candidate solution after the position update is less than the fitness of the dominant thunderstorm-level candidate solution or the tornado-level candidate solution with the current smallest fitness, the storm-level candidate solution exchanges positions with the dominant thunderstorm-level candidate solution or the tornado-level candidate solution to become the dominant candidate solution, and a new storm-level candidate solution is randomly generated.

[0157] The new storm-level candidate solution is independently subjected to boundary verification and fitness calculation, and is added to the population again from the initial stage of step S344;

[0158] S345. If the position is not exchanged, the jump operation is performed, and the expression is:

[0159]

[0160] Where l and u represent the lower and upper limits of the search region, respectively; δ2 represents the sign change; ||·|| represents the norm operator; ν is the defined exponential function;

[0161] Among them, the expression defining δ2 is:

[0162]

[0163] The expression defining ν is:

[0164]

[0165] S35. Set a tolerance level r and a tolerance threshold R according to the tolerance mechanism, and determine whether the algorithm falls into a local optimal state by using the tolerance level and the tolerance threshold; calculate the tolerance level r, which is expressed as:

[0166]

[0167] Where r is the tolerance level; is the position of the optimal solution of the previous iteration; is the position of the optimal solution of the current iteration;

[0168] When the tolerance level r is greater than or equal to the tolerance threshold R, it is determined that a local optimal solution has been reached; the random storm-level candidate solution is initialized according to differential evolution, and the tolerance level r is reset to zero, and the iterative execution returns to S33; otherwise, the termination judgment is entered;

[0169] The operation of initializing the random storm-level candidate solution is the mutation-crossover-selection operation of differential evolution, and its expression is:

[0170]

[0171] Where F is the scaling factor. If F is too small, the algorithm may fall into a local optimum, and if it is too large, it will not converge easily. Therefore, F is usually taken as a random number between [0.4, 1]. is the mutation operator individual; To randomly select three individuals with different indexes in the population, these three randomly selected individuals are mutated; CR represents the crossover probability of the algorithm, and its value is between [0,1]; j rand is a random integer used to ensure that the individuals after crossover are not all taken from the parent generation;

[0172] S36. When the maximum number of iterations reaches a preset threshold or meets the convergence condition, the optimal solution, i.e., the optimal ballast water allocation plan, is output; otherwise, the process returns to step S33 for iterative execution; the triggering rule for the convergence condition is that the tolerance level returns to zero and becomes equal to the tolerance threshold again.

[0173] The present invention has the following beneficial effects: the present invention optimizes and calculates the ship's ballast water through an improved tornado optimization algorithm, effectively avoiding the problem that traditional methods are prone to falling into local optimality, and can still efficiently search for the global optimal solution under complex multi-constraint conditions; by dynamically adjusting the algorithm convergence strategy, the number of iterations is reduced, and the timeliness requirements of real-time ship control are met while improving the calculation efficiency; multiple safety constraints are integrated into the objective function to ensure the safety indicators of the ship after ballast water adjustment.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements 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 ship ballast water regulation method based on an improved tornado optimization algorithm, characterized in that: include: S1. Acquire ship ballast tank information and current ballast water status, and construct an objective function based on the ship ballast tank information and the current ballast water status; S2. Construct ballast water constraints based on the objective function; S3. Design an improved Tornado optimization algorithm and solve the objective function in combination with the ballast water constraints to obtain the optimal solution as the current ballast water regulation plan; S4. Regulate the ballast water according to the ballast water regulation plan.

2. The ship ballast water regulation method based on the improved tornado optimization algorithm according to claim 1 is characterized in that: The expression for constructing the objective function is: Where, P is the ballast water transfer volume; Δh ji is the allocation quantity of ballast tank i; S i is the bottom area of ​​ballast tank i; ρ sw is the density of seawater; n is the number of ballast tanks; Wherein, the allocation amount Δh of the ballast compartment i is calculated ji , whose expression is: Δh ji =h ji -h (j-1)i (2) Where h ji is the water level of ballast tank i at stage j; h (j-1)i is the water level of ballast tank i at stage j-1.

3. The ship ballast water regulation method based on the improved tornado optimization algorithm according to claim 1 is characterized in that: The ballast water constraints include: hull balance constraints, ballast tank water level constraints, ballast water initial loading capacity constraints, and ballast total water capacity detection constraints.

4. The ship ballast water regulation method based on the improved tornado optimization algorithm according to claim 3 is characterized in that: The ballast water constraints are specifically: S21. The hull balance constraint condition refers to the hull maintaining moment balance under the combined action of multiple moments. Its expression is: M B +M H +M T +M R =0 (3) Where M B is the ship ballast moment; M T is the weight's pitch moment; M H is the heeling moment of the weight; M R is the restoring torque; The hull balance also satisfies the hull's longitudinal and transverse angles θ≤[θ], Among them, [θ] and are the permissible thresholds for pitch and heel angles, respectively; S22. The ballast tank water level constraint condition means that the ballast tank water level cannot be higher than the cabin height and satisfies 0≤h j ≤H j ; Among them, h j is the ballast tank water level; H j The maximum water level allowed in the ballast tank; S23. The initial ballast water loading constraint condition means that the initial ballast water loading quantity satisfies Q min ≤Q i ≤Q max Among them, Q min is the minimum ballast water loading; Q max is the maximum ballast water loading capacity; S24. The constraint condition for the detection of total ballast water volume is that the total ballast water volume must remain unchanged before and after allocation.

5. The ship ballast water regulation method based on the improved tornado optimization algorithm according to claim 1 is characterized in that: The specific process of solving the objective function by combining the improved tornado optimization algorithm with the ballast water constraint condition includes: S31. Initialize the optimization population by introducing Cat chaos mapping and reverse learning strategy; the individuals in the initialized optimization population are the water level regulation schemes for the ballast tanks; wherein the initialization process is: S311, generating a chaotic sequence according to the Cat chaotic map and mapping it to the upper and lower bounds of the variable [l j ,u j ] to obtain the initial population, the upper and lower bounds of the variables are the upper and lower bound solutions of the ballast tank capacity; S312: Calculate the reverse solution of the initial population solution according to the reverse learning strategy, and its expression is: X′ i =L i +U i -X i (4) Where, L i 、U i are the upper and lower bounds; X i is the initial population solution; X′ i is the reverse population solution; S313, merging the initial solution and the reverse population solution into a mixed solution set {X′ i ,X i }; S32. Calculate the fitness of all solutions in the mixed solution set according to the objective function and the ballast water constraint condition, and the expression is: Where P is the ballast water transfer amount obtained by the objective function; λ is the penalty function, and the expression defining the penalty function is: λ=λ0·(1+cos(πt / T) (6) Where: λ0 is the initial coefficient of the penalty function; t is the number of iterations; The mixed solution set {X′ i ,X i Sort by fitness from small to large, and select the top N candidate solutions with the smallest fitness as the final initial population; According to the fitness, the final initial population is divided into tornado-level candidate solutions, thunderstorm-level candidate solutions and storm-level candidate solutions; wherein the tornado-level candidate solutions are the first n candidates with the smallest fitness. o solutions, the thunderstorm-level candidate solution is n solutions whose fitness is greater than that of the tornado-level candidate solution. t solutions, the storm-level candidate solution is n of the remaining fitness w a solution; S33. Dynamically update the speed of the storm level candidate solution by introducing a nonlinear inertia weight. The speed is the deployment direction that determines the deployment amount of the ballast tank. The update expression is: Where i = 1, 2, ..., n w Indicates that the population size is n w Storm Index; is the new velocity vector of the candidate solution for the i-th storm level; is the current velocity vector of the candidate solution for the i-th storm level; rand is a random number uniformly distributed in the interval [0,1]; η is the convergence factor; ω1 is the nonlinear inertia weight; μ represents the fuzzy adaptive random number; R l and R r are the curvature radii of the storm tracks in the Northern Hemisphere and the Southern Hemisphere respectively; c is a random number in different ranges; define f, CF l ,CF r The expression is as follows: f=2·Ω×sin(-1+2·rand) (8) Where Ω represents the angular velocity; -1+2·rand represents the random value of the specified dimension, which is the index of the water level adjustment amount of a single cabin; Where, is the component of the pressure gradient force in the current direction of the i-th storm at the t-th iteration; S34. Adopting an adaptive external disturbance factor, reconstructing the evolution rule of the storm-level candidate solution to the tornado-level candidate solution; the evolution rule is: S341. Update the position of the storm-level candidate solution based on the speed of the storm-level candidate solution. The expression is: Where, is the next position vector of the candidate solution of storm level i at the t+1th iteration; is the current position vector of the i-th storm level candidate solution at the t-th iteration; is the current position vector of the i-th tornado-level candidate solution at the t-th iteration; The difference between the evolution of the storm-level candidate solution into the tornado-level candidate solution and the random formation of the storm-level candidate solution; rand ω is the index vector of randomly selected storms; S342. When the fitness of the storm-level candidate solution after the update is less than the fitness of the dominant thunderstorm-level candidate solution or tornado-level candidate solution with the lowest fitness, the storm-level candidate solution exchanges positions with the dominant thunderstorm-level candidate solution or tornado-level candidate solution to become the dominant candidate solution, and a new storm-level candidate solution is randomly generated. The new storm-level candidate solution is independently subjected to boundary verification and fitness calculation, and is added to the population again from the initial stage of step S342; S343. If the position is not exchanged, the jump operation is performed, and the expression is: Where l and u represent the lower and upper limits of the search region, respectively; δ2 represents the sign change; ||·|| represents the norm operator; ν is the defined exponential function; a y represents the exponential parameter; S35. Set a tolerance level r and a tolerance threshold R according to the tolerance mechanism, and determine whether the algorithm falls into a local optimal state by using the tolerance level and the tolerance threshold; calculate the tolerance level r, which is expressed as: Where r is the tolerance level; is the position of the optimal solution of the previous iteration; is the position of the optimal solution of the current iteration; When the tolerance level r is greater than or equal to the tolerance threshold R, it is determined that a local optimal solution has been reached; the random storm-level candidate solution is initialized according to differential evolution, and the tolerance level r is reset to zero, and the iterative execution returns to S33; otherwise, the termination judgment is entered; S36. When the maximum number of iterations reaches a preset threshold or meets the convergence condition, the optimal solution, i.e., the optimal ballast water allocation plan, is output; otherwise, the process returns to step S33 for iterative execution; the triggering rule for the convergence condition is that the tolerance level returns to zero and becomes equal to the tolerance threshold again.

Images