CN115730700B - Self-adaptive multi-target task planning method, system and equipment based on reference points - Google Patents

Abstract

The application relates to a self-adaptive multi-objective task planning method, a system and equipment based on a reference point, wherein the method comprises the following steps: acquiring known task parameters, calling a constructed amphibious helicopter training task planning model, calculating the population scale, constructing an initial population, and performing crossover operation, mutation operation and local optimizing search operation on a parent population to generate a child population; and merging the generated child population and the parent population to calculate a standardized solution set, and generating a reference point set and a corresponding set of interest solution sets on the standardized solution set by using a reference point self-adaptive generation method based on population distribution. A new generation population is constructed by using a clustering method based on an interior angle measurement method and a scaling sorting method; and (3) performing iterative optimization until the iteration number of the population reaches the maximum iteration number, and outputting amphibious helicopter training task planning scheme data. The rapid planning of the amphibious helicopter training task is realized more efficiently.

Description

Self-adaptive multi-target task planning method, system and equipment based on reference points

Technical Field

The invention belongs to the technical field of task planning, and relates to a self-adaptive multi-target task planning method, system and equipment based on a reference point.

Background

In an amphibious task, a helicopter taking an offshore forward base as a lifting platform (LHD) can overcome the obstacle on the edge of a body of water beach by virtue of the characteristics of flexible deployment and concealment, and deliver task resources deep in a landing field simple and plain and form local advantages simultaneously, so that the abrupt and quick resolution of the amphibious task is achieved. However, the helicopter has natural defects of complex guarantee, low running strength, limited carrying capacity, easy interference and the like, so that the helicopter has more use restrictions and larger task risks. Therefore, the task of the amphibious helicopter needs to be finely planned, so that the advantages of the amphibious helicopter are fully exerted, and meanwhile, the loss is reduced as much as possible.

Currently, mission planning studies on aircraft consist are mainly focused on the field of UAV (Unmanned AERIAL VEHICLE) trunked action planning, and can be roughly divided into two parts, namely mission allocation studies and routing studies. The content of task allocation research mainly comprises allocation decisions of tasks among different resources and platforms and optimal scheduling of task execution sequences, and currently, a model which is mature in the field comprises a travel business problem (TSP) model, a Vehicle Routing Problem (VRP) model and a mixed integer linear programming problem (MILP) model. The conventional training task planning model usually takes the maximum achievement, the minimum loss, the task completion in the shortest time, the most full utilization of human resources and the like as optimization targets.

In reality, most of the task planning problems are Multi-objective optimization problems (MOP-objective Optimization Problem). At this stage, the Multi-objective evolutionary algorithm (Multi-objective Evolutionary Algorithm, MOEA) has proven to be the most effective way to solve MOP, and according to the solution idea, it can be roughly divided into three categories, pareto-based dominant relations, index-based and decomposition-based. However, the conventional training task planning method cannot directly solve the technical problem of rapid planning of the training task of the amphibious helicopter in reality.

Disclosure of Invention

Aiming at the problems in the traditional method, the invention provides a self-adaptive multi-target task planning method based on a reference point, a self-adaptive multi-target task planning system based on the reference point and computer equipment, which can efficiently and effectively realize the rapid planning of the training task of the amphibious helicopter.

In order to achieve the above object, the embodiment of the present invention adopts the following technical scheme:

in one aspect, a method for adaptive multi-objective task planning based on reference points is provided, including the steps of:

acquiring known task parameters; the task parameters comprise the total number of platforms, the total wave number of helicopter running, the dimension of solution space, the simplex segmentation parameter and the maximum iteration number;

Calling a constructed amphibious helicopter training task planning model; constraint conditions of the amphibious helicopter training task planning model comprise vertical login team quantity constraint, helicopter quantity constraint, platform guarantee resource constraint, task stage time sequence constraint based on execution sequence and task stage time sequence constraint based on platform deck scheduling, and optimization objective functions of the amphibious helicopter training task planning model comprise time-consuming objective functions of a minimum login stage, a minimum manpower loss objective function and a minimum ground threat objective function of the helicopter;

Calculating a population scale for an amphibious helicopter training task planning model based on task parameters and constructing an initial population;

Performing crossover operation, mutation operation and local optimization search operation on the parent population to generate a child population;

Combining the generated child population with the parent population to obtain a combined population, updating IDEAL points and NADIR points according to elements in a solution set corresponding to the combined population, and calculating according to the IDEAL points and NADIR points to obtain a standardized solution set;

generating a reference point set and a corresponding set of interest solution sets on the standardized solution set by using a reference point self-adaptive generation method based on population distribution;

utilizing a clustering method based on an interior angle measurement method to distribute each individual in the standardized solution set to the interest base solution set, carrying out non-dominant sorting operation in each interest base solution set based on a scaled sorting method, and selecting individuals corresponding to elements ranked at the front to form a new generation population;

and returning to the step of executing the cross operation, the mutation operation and the local optimizing search operation on the parent population to generate the child population, and outputting amphibious helicopter training task planning scheme data when the iteration number of the population reaches the maximum iteration number.

In another aspect, there is also provided a reference point-based adaptive multi-objective mission planning system, including:

The parameter acquisition module is used for acquiring known task parameters; the task parameters comprise the total number of platforms, the total wave number of helicopter running, the dimension of solution space, the simplex segmentation parameter and the maximum iteration number;

the model calling module is used for calling the constructed amphibious helicopter training task planning model; constraint conditions of the amphibious helicopter training task planning model comprise vertical login team quantity constraint, helicopter quantity constraint, platform guarantee resource constraint, task stage time sequence constraint based on execution sequence and task stage time sequence constraint based on platform deck scheduling, and optimization objective functions of the amphibious helicopter training task planning model comprise time-consuming objective functions of a minimum login stage, a minimum manpower loss objective function and a minimum ground threat objective function of the helicopter;

The population construction module is used for calculating the population scale of the amphibious helicopter training task planning model based on the task parameters and constructing an initial population;

the offspring generation module is used for executing crossover operation, mutation operation and local optimization search operation on the parent population to generate offspring population;

the merging calculation module is used for merging the generated child population and the parent population to obtain a merged population, updating IDEAL points and NADIR points according to elements in a solution set corresponding to the merged population, and calculating according to the IDEAL points and the NADIR points to obtain a standardized solution set;

the set generation module is used for generating a reference point set and a set of corresponding interest solution sets on the standardized solution set by using a reference point self-adaptive generation method based on population distribution;

The new population module is used for distributing each individual in the standardized solution set to the interest base solution set by using a clustering method based on an interior angle measurement method, performing non-dominant sorting operation in each interest base solution set based on a scaled sorting method, and selecting individuals corresponding to the elements ranked at the front to form a new generation population;

And the iteration control module is used for outputting amphibious helicopter training task planning scheme data when the population iteration number reaches the maximum iteration number.

In yet another aspect, a computer device is provided, including a memory storing a computer program and a processor implementing the steps of the adaptive multi-objective task planning method based on reference points described above when the processor executes the computer program.

In yet another aspect, a computer readable storage medium is provided, on which a computer program is stored, which when executed by a processor implements the steps of the adaptive multi-objective task planning method based on reference points described above.

One of the above technical solutions has the following advantages and beneficial effects:

According to the self-adaptive multi-objective task planning method, system and equipment based on the reference points, through extracting and analyzing the main task mode and characteristics of the helicopter in amphibious landing actions, an amphibious helicopter training task planning model is provided on the basis of reasonably supposing a real landing training scene, and the model is expressed as a multi-objective optimization model, and the aims of minimizing landing action time consumption, own personal casualties and blue-side air defense fire threat of the own helicopter are achieved. The amphibious helicopter training task planning model considers the specific reality constraint of two aspects of training resources which can be used by multiple platforms, training and guaranteeing the full-link task stage time sequence. The adaptive generation method of the reference points based on the population distribution is applied to the population sorting selection mechanism level, the optimizing capability of the irregular population is enhanced more effectively on the premise of not increasing the computational complexity, new initial solution generation, intersection, variation and local optimizing methods are provided to the population updating iteration mechanism level, the searching efficiency of the coding definition space is improved better, good performance is achieved in optimizing efficiency and maintaining population diversity, and therefore the purpose of fast planning of the training task of the amphibious helicopter is achieved more efficiently and effectively.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments or the conventional techniques of the present application, the drawings required for the descriptions of the embodiments or the conventional techniques will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to the drawings without inventive effort for those skilled in the art.

FIG. 1 is a flow chart of a reference point-based adaptive multi-objective mission planning method in one embodiment;

FIG. 2 is a schematic diagram of the flight mission and deck preparation phases of a helicopter formation in one embodiment;

FIG. 3 is a flow diagram of iterative optimization solution to a model in one embodiment;

FIG. 4 is a schematic diagram of an example of chromosome coding in one embodiment;

FIG. 5 is a schematic diagram of an example of a helicopter play plan in one embodiment;

FIG. 6 is a schematic diagram of a process for crossing one chromosomal code with another chromosomal code according to one embodiment;

FIG. 7 is a hyperplane schematic constructed with the point closest to the coordinate axis as the Nadir point;

FIG. 8 is a hyperplane diagram constructed with the point of maximum target on each coordinate axis as the Nadir point;

FIG. 9 is a schematic diagram of a result obtained by clustering the normalized solution sets in one embodiment;

FIG. 10 is a schematic diagram of a constructed pseudo-linear hyperplane and a set of reference points in one embodiment, where (a) is the constructed pseudo-linear hyperplane and (b) is the set of reference points generated;

FIG. 11 is a schematic diagram of an embodiment in which the polar space Γ ³ is partitioned;

FIG. 12 is a schematic diagram of a result obtained by clustering the normalized solution sets according to another embodiment;

FIG. 13 is a schematic diagram of the modular composition of an adaptive multi-objective mission planning system based on reference points, in one embodiment.

Detailed Description

The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

It is noted that reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.

Those skilled in the art will appreciate that the described embodiments of the application may be combined with other embodiments. The term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.

And (3) training task analysis: the helicopter grouping consisting of the general helicopter and the special helicopter can execute various tasks including vertical delivery of personnel, equipment and materials, near-distance air support, low-altitude empty weight capturing, impact and anti-diving mine sweeping on sea surface targets and the like in amphibious landing tasks. In the amphibious task process, a lift taking off from the LHD can vertically land at the depth of the target coast simple and plain, and the delivery machine can landing on a plane in cooperation with manpower to assist landing team members to rapidly advance on the target coast or to implement assault on the side wings or rear main points of the target so as to achieve the local task target. Because the helicopter is limited by the intensity of the movement and the load at the present stage, only a light task unit can be delivered on a small scale, and the vertical landing unit mainly bears the cooperative task of the plane landing unit in a large-scale amphibious landing task.

According to the analysis, the vertical landing task can be abstracted into the distribution of the load (the landing manpower) of the general helicopter on the offshore platform to different points of the target, and the method specifically comprises the selection and calculation of the points of assault, the moment of assault, the landing field, the control point of the helicopter navigation and other elements. The execution main body of the vertical landing task is an offshore platform general helicopter, and the vertical landing task also comprises an offshore platform special helicopter for providing accompanying navigation, reconnaissance and shielding.

In the implementation process of the vertical landing task, the effects of the LHD and the ship-to-ship cooperation of the helicopter, the cooperation of the helicopter and the helicopter in the formation of the helicopter, the cooperation of the general helicopter and the special helicopter in different wave passes, the cooperation of the formation of the task groups of different targets in the same wave pass, the effect of the cooperation of the vertical landing unit and the work species of the plane landing unit on the efficiency of the vertical landing task and the like are considered, and the effects of the platform guarantee capability on the starting efficiency of the helicopter, the effects of the load, the speed and the course of the general helicopter on the manpower delivery efficiency, the effects of the information weight, the air-making right and the air-preventing pressing effect on the targets on the flight safety of the helicopter and the like are specifically included.

The special helicopter for the offshore platform can realize the flying height far lower than that of a fixed wing aircraft and the maneuvering speed of far superwheel type or caterpillar equipment by virtue of the special flying characteristics. Therefore, in the near-distance air support task implemented by taking the helicopter special for the offshore platform as a main body, the special helicopter marshalling taking off from the LHD deck or standing on the front line airspace can be quickly maneuvered to the target airspace by utilizing ultra-low-altitude flight concealment, and accurate target assault is implemented on important targets such as target vehicles, radar stations, target battle areas, communication hubs, leading sentry posts, simple work, target points, surface ships and ground units, and the like, so that the helicopter marshalling has irreplaceable important roles.

Similar to a vertical landing task, a near-distance air support task can be abstractly understood to be the distribution of a helicopter load special for an offshore platform to different targets, and specifically comprises the selection and calculation of factors such as a task target, a assault time, a helicopter route control point and the like. The main implementation subject of the near-distance air support task is an offshore platform special helicopter, and can also comprise an unmanned plane for providing job site situation reconnaissance and the like. In the task implementation process, the influence of factors such as ship surface guarantee capacity, special helicopter load, speed, voyage, air defense targets, operation field situations and the like on the hitting efficiency of the helicopter targets is also required to be considered.

In one embodiment, as shown in fig. 1, the present application provides a reference point-based adaptive multi-objective task planning method, which includes steps S11 to S18:

S11, acquiring known task parameters; the task parameters include the total number of platforms, the total wave number of helicopter play, the solution space dimension, the simplex segmentation parameters and the maximum iteration number.

It will be appreciated that, before the optimization calculation begins, predefined relevant task coefficients, such as, but not limited to, the total number of platforms carrying the lift-off platform of the helicopter and the total number of wave times the helicopter is lifted, may be obtained, as well as relevant known parameters of the optimization calculation, such as solution space dimensions, simplex segmentation parameters, and maximum number of iterations.

S12, calling a constructed amphibious helicopter training task planning model; constraint conditions of the amphibious helicopter training task planning model comprise vertical landing team quantity constraint, helicopter quantity constraint, platform guarantee resource constraint, task stage time sequence constraint based on execution sequence and task stage time sequence constraint based on platform deck scheduling, and optimization objective functions of the amphibious helicopter training task planning model comprise time-consuming objective functions of a minimum landing stage, a minimum manpower loss objective function and a minimum ground threat objective function of the helicopter.

It will be appreciated that amphibious joint logging is the most complex form of action, and that the training mission planning model involves a plethora of elements, and that appropriate normative assumptions are necessary to make the model relatively simple. For convenience of description, the attacking party among the attacking and defending parties can be called red party, and the defending party can be called blue party. Model hypothesis 1 was: the method comprises the steps that each training task in the red landing action is completed by a plane landing team, a vertical landing team and a helicopter, wherein the plane landing team training task is formulated by a pre-action staff plan, and the helicopter mainly bears the vertical landing task and the close-range support task.

Model hypothesis 2 is: the vertical landing tasks are completed by the general helicopter and the special helicopter together, the type and the number of vertical landing teams delivered to each target of the blue party and the number of the two types of helicopters required are formulated in a pre-action staff plan, the short-distance support tasks are completed by the special helicopter only, and each short-distance support task grouping consists of the special helicopters with fixed number.

Model hypothesis 3 is: the helicopter cannot influence the training tasks executed by the helicopter due to faults or damages, the LHD cannot be attacked, sufficient oil and ammunition supply and various service guarantees can be stably provided for the helicopter, and in addition, a command information system of a landing team can provide stable and effective command and communication guarantees for all tasks.

Model hypothesis 4 was: the situation that the LHD is close to the landing beach (such as 50 km) is selected for research, and the time for carrying out the task by the helicopter task grouping and the lift-off is short, so that the helicopter is assumed to take double-wave cyclic lift-off, and the helicopter grouping with the same task is lifted off at the same or nearby wave as much as possible.

Model hypothesis 5 was: each helicopter is assumed to have the same ship surface dispatching, machine service guarantee, take-off departure, formation flying (between LHD and landing zone) and landing time, the helicopter is assumed to fly at an ultra-low altitude at a uniform speed in the landing zone, the flying time is dependent on the flying path length of the helicopter in the landing zone, and in addition, each vertical landing task group and short-distance support task group are assumed to have the same and fixed landing time and ground striking time respectively.

Model-related variable definition: the relevant variable symbols in the model are given in table 1.

TABLE 1

Model problem constraints include two major classes, namely, training resource constraints and task phase timing constraints, wherein the training resource constraints include:

(1) Vertical log-in team number constraint: according to model assumption 4, in order to deliver more vertical login queues simultaneously in one wave-time, the vertical login queues should be distributed and deployed on different LHDs, and since the total number of vertical login queues that can be put into by red parties in login actions is limited, the number of vertical login tasks that each LHD can support must also meet the limit of the number of vertical login queues that the LHD loads, specifically, The following constraints should be satisfied:

Meanwhile, the total number of all vertical login teams scheduled for delivery of all vertical login tasks in one login action also meets the following constraint:

(2) Helicopter quantity constraint: according to model assumption 2, the sum of certain types of helicopters required by task grouping of helicopters circularly driven by adjacent wave times on a certain LHD is limited by the total number of the helicopters carried by the LHD:

In formula (3), the number of dedicated helicopters required for each individual training task And number of general helicoptersThe ratio between the two is determined by the task type c, when c=2 the helicopter consist performs the short-range support task, the task consist consists of only dedicated helicopters, therefore the number of general helicoptersWhen c=1, the helicopter consist performs a vertical landing task, the task consist consists of a general helicopter and a special helicopter providing it with accompanying shelter, usually the number of special helicoptersShould not be less than the number of general helicopters1/2 Of (a), it is therefore assumed in the present application that:

In the above formula (4) Representing the number of generic helicopters needed for each vertical landing mission, in addition to rounding up xDepending on the number of various vertical login teams required to perform an air assault on Lan Fangdi p nodes in the project of the party before the red party acts

(3) LHD guarantee resource constraints: according to model hypothesis 4, limited to the number of take-off positions on each LHD flight deck, the number of helicopters grouped for a helicopter mission that is driven at the same wave number should satisfy the following equation (6):

in addition, the helicopter group allocated by the same training task should start as much as possible in the same wave time or adjacent wave time, thus the formula The constraints should be satisfied:

In the above formula (7), n _w represents the minimum number of waves required for all helicopter groups allocated to the class c training task of the target p to all play, i.e. the sum of the number of two classes of helicopters required for the task group divided by the upward rounding of the quotient of all take-off positions of all LHDs of the red side.

The task phase timing constraints include:

(1) Task phase timing constraints based on execution order: the complete helicopter training task consists of 9 parts, namely a ship surface dispatching stage, a machine service guarantee stage, a take-off departure stage, a formation flight stage (cruising) from an LHD to a coastline, a formation flight stage (ultra-low altitude burst prevention) from the coastline to a target node, a task execution stage, a formation flight stage (ultra-low altitude burst prevention) from the target node to the coastline, a formation flight (cruising) from the coastline to the LHD, a landing ship approach stage and the like, and the following constraint is easily satisfied between the starting time and the ending time of each stage in each independent helicopter training task:

In the formula (8), the amino acid sequence of the compound, AndThe following relationship is satisfied:

in (9) Representing the time spent in the kth sub-stage of performing a class c training task on the p node target of the landing field at the w-th wave play at the red LHD. Wherein, the time consumption of the ship surface transferring stage, the machine service support stage, the take-off and departure stage and the approach ship landing stageAndFor an assumed fixed value, the task execution phase is time-consumingDepending on the type of mission, the helicopter formation time of flightAndAnd dividing the planned optimal path length by the helicopter formation flying speed.

Notably, while the present application includes 9 sub-phases for each helicopter training mission, not all helicopter consist assigned to a training mission will go through the 9 phases entirely. For example, in the ship surface dispatching of stage 1, if all task groups of w and w+2 wave numbers circularly carried out on a certain LHD need the same total number of helicopters, the helicopter carrying out w wave operation can directly carry out the maintenance of the service and service on a take-off position after landing, and then directly lift off the task groups forming w+2 wave numbers to execute training tasks. Thus in formula (9)The following formula is satisfied.

T ₁ in equation (10) represents the assumed ship surface deployment time.

(2) Task phase timing constraints based on LHD deck scheduling: due to limited space of the LHD flight deck, the helicopter usually occupies a take-off station for service support on the LHD, so that a helicopter formation ready for landing must wait for the LHD flight deck to empty (i.e., after the next helicopter takes off) before landing. Thus, for a helicopter formation to take off at 2 adjacent wave times on the same LHD, its take-off and landing moments must satisfy the following constraints:

In the formula (11), the amino acid sequence of the compound, Representing the landing time of a helicopter group which performs a class c training task on a p node target of a landing field at the w' th wave play on the first LHD of the red party,The take-off times of helicopter consist that perform the same task at the w "th wave order of play on the same LHD are represented.

As shown in fig. 2, if it is assumed that the helicopter task consist that starts at the same wave at all LHDs of the red party must take off simultaneously, after all the helicopter consists of the w th wave fall, the ① No. helicopter consist with shorter preparation time for performing the service guarantee directly at the take-off position will wait for the ② No. helicopter consist that needs to be deck-transferred to replace the helicopter to take off together after completing the service guarantee. After ① and ② helicopter groups are all lifted off, the helicopter groups ③ and ④ which return to the voyage after the w-1 wave is completed in task can only land on the emptied deck.

Model optimization objective: on one hand, because the beach head can accommodate limited manpower, the red party landline can only throw in one by one, in order to prevent the blue party from possibly initiating reverse impact at any time, the first wave landline must seize and consolidate landing sites rapidly, and ensure the safe and rapid landing of the subsequent team; on the other hand, blue squares facing to support firm work are disadvantageous in terms of both fire and protection due to lack of heavy equipment by the red square head wave land-on team. Therefore, the red team must reasonably plan the manpower application, and reduce the casualties as much as possible on the premise of ensuring that the scheduled training task is completed on time. By integrating the analysis, the optimization targets of the model, such as time consumption, manpower loss, ground threat of the helicopter and the like, of the training in the red party open landing field landing stage are selected.

(1) Minimizing login phase time consumption: in amphibious landing actions, a defender can organize maneuvering manpower to perform reverse impact when the foothold of an attacking party is unstable, whether the defender can quickly take up and effectively consolidate landing sites before initiating reverse impact is an important factor for determining whether one landing action is successful or not. Therefore, the model selects the total time consumption of all training teams of red parties to complete all established tasks in the open login field stage as a time consumption objective function in the login stage:

z³＝T_fin-T₀ (12)

In the formula (12), T _fin represents the finishing time of the login stage, namely the time when the red square head wave land team finishes various tasks required for opening and consolidating the beach login place; t ₀ corresponds to the login action initiating time, namely the take-off starting time of the helicopter with the first wave of the red square.

(2) Minimizing human power loss: the red party manpower loss comprises plane landing manpower loss and vertical landing manpower loss, and the manpower loss objective function can be specifically expressed as:

In the formula (13), the amino acid sequence of the compound, The number of manpower lost by all training teams of red party trained at the p-th blue party node in unit time at the time t is represented,The challenge start time at the p-th blue node is indicated, and T ^p is indicated for the challenge duration in minutes. The application utilizes the existing Lanchester equation to calculate the manpower loss of both red and blue.

(3) Minimizing the ground threat to helicopters: in order to avoid the model being too complex, the influence of manpower grouping of two training tasks and subsequent available helicopters on the LHD after the damage of the red Fang Zhi helicopter is not considered, and the loss of the helicopter is quantitatively represented only by the accumulation of the blue party anti-air threat in the helicopter navigation process:

In the formula (14), k ^Scale represents a scale factor, AndE (x, y, t) represents the sum of all ground threat values of a blue party received by each helicopter in a unit time at a coordinate (x, y) at the time t. The application properly simplifies the calculation of E (x, y, t), and only considers the air-fire resistance coefficient at each node of the blue square, the distance between a red Fang Zhi machine and the node and the shielding effect of the terrain on blue Fang Fangkong fire, and E (x, y, t) can be calculated according to a formula (15).

Η _SAM in the formula (15) is a boolean variable indicating whether or not the node p and the coordinates (x, y) are occluded by the terrain, 1 when not occluded and 0 when occluded; The air defense fire coefficient at the blue square node p is represented; d (x, y, p) represents the planar distance between node p and coordinates (x, y). the position (x, y) of each helicopter at the time t is obtained by solving a path planning based on dijkstra algorithm in the decoding operation.

S13, calculating a population scale for the amphibious helicopter training task planning model based on task parameters and constructing an initial population;

s14, performing cross operation, mutation operation and local optimization search operation on the parent population to generate a child population;

S15, merging the generated child population and the parent population to obtain a merged population, updating IDEAL points and NADIR points according to elements in a solution set corresponding to the merged population, and calculating according to the IDEAL points and the NADIR points to obtain a standardized solution set;

s16, generating a reference point set and a corresponding set of interest solution sets on the standardized solution set by using a reference point self-adaptive generation method based on population distribution;

s17, distributing each individual in the standardized solution set to the interest solution set by using a clustering method based on an interior angle measurement method, performing non-dominant sorting operation in each interest solution set based on a scaled sorting method, and selecting individuals corresponding to elements ranked at the front to form a new generation population;

S18, returning to the step of executing the cross operation, the mutation operation and the local optimizing search operation on the parent population to generate the child population, and outputting amphibious helicopter training task planning scheme data until the iteration number of the population reaches the maximum iteration number.

It can be appreciated that conventional planning methods are inefficient in solving problems where the objective function and constraints are nonlinear, and are sensitive to weights and optimization order. Along with the continuous development of the multi-objective optimization problem towards reality and complexity, the multi-objective evolutionary algorithm becomes a main means for solving the problem.

According to the self-adaptive multi-objective task planning method based on the reference points, the main task mode and the characteristics of the helicopter in the amphibious landing actions are extracted and analyzed, the amphibious helicopter training task planning model is provided on the basis of reasonably supposing a real landing training scene, the model is expressed as a multi-objective optimization model, and the aims of minimizing the time consumption of landing actions, casualties of a host person and the air fire threat of the host helicopter to the blue side are achieved. The amphibious helicopter training task planning model considers the specific reality constraint of two aspects of training resources which can be used by multiple platforms, training and guaranteeing the full-link task stage time sequence. The adaptive generation method of the reference points based on the population distribution is applied to the population sorting selection mechanism level, the optimizing capability of the irregular population is enhanced more effectively on the premise of not increasing the computational complexity, new initial solution generation, intersection, variation and local optimizing methods are provided to the population updating iteration mechanism level, the searching efficiency of the coding definition space is improved better, good performance is achieved in optimizing efficiency and maintaining population diversity, and therefore the purpose of fast planning of the training task of the amphibious helicopter is achieved more efficiently and effectively.

In order to express the sequence relation among the training tasks as simply and clearly as possible, the application adopts a discrete task list coding form similar to VRP. According to model constraint 2, when the total number of LHDs and total wave numbers of running are given, the total number of helicopters required by all the training tasks in the chromosome coding is also determined, so that the training tasks requiring different numbers of helicopters in the amphibious helicopter training task planning model can certainly lead to different lengths of the chromosome coding. In addition, according to the analysis, all training tasks are difficult to be contained in the corresponding chromosome codes of any decision variable of the amphibious helicopter training task planning model, so that the traditional genetic operator has difficulty in realizing the crossing operation of chromosomes with different lengths, and is more difficult to effectively search the definition domain of the problem by virtue of the chromosomes only containing limited coding information. Aiming at the problems, the application improves the population updating iterative mechanism of the evolutionary algorithm, redesigns the cross operation and variation operation methods applicable to different length codes, and creatively proposes an initial solution generating method and a local searching method to improve the optimizing efficiency of algorithm processing.

Known parameters include the solution space dimension (denoted m), the simplex segmentation parameter (denoted H) and the maximum number of iterations of the algorithm (denoted g _Iter). First, the population size |p _g | is calculated according to equation (16):

then constructing an initial population P ₀ by an initial population generation method (Initial Population Generation Method based on Traversing 1st-Class Neighborhoods, IPGM-T1N) based on searching the class 1 neighborhood; step two, performing crossover, mutation and local optimization operations on the parent population P _g to generate a child population o' _g+1; third, combining the generated o '_g+1 with the parent population P _g to obtain a combined population P' _g+1, updating the parameters z _Ideal and z _Nadir according to the elements in the solution set z '_g+1 corresponding to the combined population P' _g+1, and further calculating according to the formula (17) to obtain a standardized solution set

M in the above equation (17) represents the solution space dimension,AndRepresenting the original solution set z and the normalized solution set, respectivelyIs a subject of the group (B). Fourth step, inBased on the above, the existing reference point generation method based on population distribution is utilized to obtain a reference point set Λ and a set Ω of corresponding interest solution sets, wherein Λ= { λ ₁,…,λ_n},Ω＝{ω₁,…,ω_n},n＝|P_g |. Fifthly, the conventional clustering method based on the interior angle measurement (INNER ANGLE Measure, IAM) is used for collecting the standardized solutionsEach individual of (3)Assigned to the Liji solution set omega _λ epsilon omega, non-dominant ranking operations are performed in each omega _λ based on the existing scaled ranking methods (Scalarization-based Sorting Method, SSM), and individuals corresponding to the top-ranked elements are selected to form a new generation population P _g+1. And finally, repeating the second step to the fifth step until the iteration number of the population reaches the maximum iteration number g _Iter, and outputting the amphibious helicopter training task planning scheme data obtained through final optimization.

In one embodiment, several principal component relationships of the proposed model optimization solution algorithm are shown in FIG. 3. The chromosome coding mode adopted in the population iteration process is as follows: decomposing all training tasks meeting the condition of setting the number of helicopter groups into a single task list with the number of helicopter groups equal to 1 and adopting discrete double-row chromosome coding; adding a vacancy gene in the chromosome coding of the task list,

Specifically, as shown in FIG. 4, all conditions for setting the number of helicopter groupings are satisfied (i.e.) Is decomposed into the number of helicopter groupsAdopts discrete double-row chromosome coding:

In the formula (18), σ _i represents a chromosome-encoding gene, and the meanings of the first row c and the second row p are the same as those in table 1; the chromosome length, i.e. the total number of helicopter training tasks, is represented.

According to the constraint of formula (3), if the total number of helicopters of a certain type on the LHD is insufficient to guarantee continuous invocation of training tasks of two adjacent wave passes, limited helicopter resources must be allocated to a certain wave pass, and the other wave pass can only schedule tasks requiring other types of helicopters or not schedule tasks. Therefore, it is necessary to encodeTo indicate the idle state of some take-off bits of a certain LHD in a certain wavelength, in the present applicationTo represent such an idle state and assuming that the number of idle takeoff bits is the maximum common factor for the number of helicopters required for all training tasks.

According to model hypothesis 2, the neighborhood in which each gene in the chromosome coding is located is classified into 2 classes: genes corresponding to vertical login tasks (hereinafter collectively referred to as class 1 training tasks) requiring variable helicopter numbers are located in class 1 neighborhoods; genes corresponding to short-distance support tasks (hereinafter collectively referred to as class 2 training tasks) requiring a fixed number of helicopters are located in class 2 neighborhood. It is readily apparent that the variation in chromosome length is caused by a variation in class 1 training tasks in the code.

The chromosome decoding mode adopted in the population iteration process is as follows: converting the chromosome codes of the task list into a play list representing the play plan of each helicopter; obtaining a starting time list of a vertical login task and a close-range support task for each target through the play list; according to the play list, the initiation time list and the preset training plan of the plane login team, calculating the change of the weapon forces of all target nodes of the login field along with time by utilizing an improved Lanchester model, obtaining a state vector of the nodes of the login field, deducing the progress direction of login actions through a state transfer model of the login field, and determining the achievement condition of the training purpose and the training task efficiency index set.

It will be appreciated that with respect to the decoding portion of the algorithm, the input parameters for a particular decoding operation are a list of discrete tasksThe decoding process is roughly as follows: first, discrete codes characterizing training task sequencesConversion to a list of movements characterizing each helicopter movement planBy passing throughThe obtained initiation time list of the vertical login task and the close-range support task of each target of the red party to the blue partyCombining with a scheduled training plan of a plane landing team, calculating the time-dependent change condition of manpower of both parties against each target node of a landing field through a Lanchester model improved in the art, further obtaining a landing field node state vector delta, deducing the progress direction of landing actions through a landing field situation transfer model, and finally determining whether the red party can finally achieve the action purpose and training task efficiency index set

In one embodiment, further, the process of converting to a play list includes the following:

Calculating the total number of vertical login teams delivered by each vertical login task in the chromosome coding of the task list and the number of various helicopters required by each vertical login task and a close-range support task;

Selecting 1 class of training tasks meeting the training resource constraint in a task list according to the total vertical login team quantity and the number of various helicopters, assigning values for elements in a play list according to a set assignment formula, and adding the executed tasks to an executed task list according to the original list sequence; the class 1 training task is a vertical login task;

After all the class 1 training tasks meeting the training resource constraint are inserted into the play list, selecting class 2 training tasks meeting the training resource constraint in the task list according to the total number of vertical login teams and the number of various helicopters, and correspondingly updating the play list and the executed task list; class 2 training tasks are close-range support tasks

In particular, inAndIs to solve the stage: first, the chromosome coding of the task list is calculatedThe total number u _i of vertical log-in teams (Airborne Landing Units, ALUs) delivered for each vertical log-in task, and the number hi of each type of helicopter required for each vertical log-in task and close-range support task. Second, selecting a training task list according to u _i and hiClass 1 training tasks satisfying the constraints of formulas (1) - (7) are given according to a set assignment formula (19)Assigned values of elements in the list and adding the executed tasks to the list in the original list orderIs a kind of medium. Thirdly, inserting all class 1 training tasks meeting the constraintThen searching according to the same principle as the second step of the sectionClass 2 training tasks meeting constraints in the system and updating the same accordinglyAnd

Assuming that there are 3 LHDs on the red side and the number of helicopters is sufficient, the 3 wave times are accumulated, the chromosome encoding in fig. 4Can be transcribed into a helicopter play schedule as shown in figure 5 Is a row of w lines and is arranged in the row,A matrix of columns is provided which,The task executed by the helicopter taking off from the j take-off position of the first LHD in the w-th wave is represented as the value shown in a formula (19):

further, initiate time list May include the following:

drawing a minimum risk route set from the helicopter to each task target by using Dijiestra algorithm according to the task targets of each training task in the play list and the distribution condition of target air defense nodes in a landing field;

And solving and obtaining an initiating time list of the task execution stage of each training task by using the play list and the route set.

In particular, inIs to solve the stage: first step, according to the helicopter's movement planThe target p _i of each training task in the system is combined with the known or assumed distribution situation of target air defense firepower in a landing field, dijkstra (Dijkstra) algorithm rule is utilized to draw a minimum risk route set from the helicopter to each task target p _i Second, solving a list of starting moments (starting moment of the vertical landing team for the vertical landing task and starting moment of the ground attack of the special helicopter for the short-distance support task) of the task execution stage of each training task

Further, the process of determining the training task performance index set may specifically include the following processes:

setting the state vector value of the corresponding landing field node at the moment of initiating the landing action to be zero;

solving the bearing risk in the flight process of each helicopter grouping task based on the setting;

according to a starting time list and a preset training plan of a plane login team, finding out a login field node which is in antagonism in each simulation time, calculating manpower loss and time consumption of each login field node against both sides, updating a corresponding login field node state vector according to login field situation transfer conditions, and carrying out numerical integration operation on the manpower force vectors of both sides at each target node by setting simulation step length until the login field node state vector meets a preset target of login action during a login field stage or the training target cannot be achieved;

And (3) counting total consumed time in the open-up landing stage, accumulated manpower loss of the vertical landing team and the plane landing team, and bearing risks of all helicopters executing the training tasks to form a training task efficiency index set.

In particular, inIs to solve the stage: first, assume that at time T ₀, where login actions are initiated, forLanding field node state vector delta (p) =0. Secondly, solving the bearing risk of each helicopter marshalling task of the red party in the flight process; third step, according to each helicopter training task initiating timeAnd (3) combining a preset training plan of a plane login team, finding out a login field node where antagonism is occurring in each simulation moment, calculating the manpower loss and the time consumption of antagonism (namely, the time consumption) of each node in the antagonism of both sides by using formulas (12) - (15), and carrying out numerical integral operation on the force vector of the manpower of both sides in each target node by taking t _step as a set simulation step length so as to meet the requirement of a preset target in the red side login action opening and login field stage or determining that the action target cannot be achieved by the red side. Finally, the total consumed time z ¹ at the open landing field stage, the accumulated manpower loss z ² of the red square vertical landing team and the plane landing team and the bearing risk z ³ of all helicopters for executing training tasks are calculated to form a target vector

In one embodiment, regarding the process of constructing the initial population in the step S13, the following processing steps may be specifically included:

searching all 1 class training task combinations meeting the number constraint of vertical login teams to form a1 class training task set;

Determining the total task number contained in the target chromosome according to the number constraint of the helicopters; the target chromosome is a chromosome containing elements in the class 1 training task set;

Constructing an arrangement set containing all arrangement forms of the corresponding elements in the class 1 training task set in a chromosome with the length of the total task number based on the total task number and the task number contained in the corresponding elements in the class 1 training task set;

For each element in the arrangement set, taking the residual space in the chromosome corresponding to the element as a class 2 neighborhood and carrying out random assignment to obtain an arrangement complete code; the class 2 neighborhood is the neighborhood where the class 2 training task corresponding genes are located;

Chromosome decoding is carried out on the arrangement complete codes to obtain corresponding arrangement target values;

If the permutation target value is a feasible solution, combining the permutation complete code and the corresponding permutation target value as an individual in the initial population;

Returning to the step of determining the total task number contained in the target chromosome according to the helicopter quantity constraint until each task contained in each element in the permutation collection is traversed, and outputting an initial population containing an initial coding collection and an initial solution collection.

It can be understood that, since the information containing all 1 class training tasks cannot be guaranteed in a single chromosome code, and different 1 class training tasks in each chromosome directly determine the difference of the coding lengths, in order to make the algorithm fully search 1 class neighborhood in the feasible domain, and also make each chromosome code in the initial population contain more training tasks as much as possible while meeting the constraints, an initial population generation method (Initial Population Generation Method based on Traversing st-Class Neighborhoods, IPGM-T1N) based on searching 1 class neighborhood is designed, a feasible coding set containing all different 1 class training task permutation and combination (i.e. all possible lengths) is generated, and is used as the initial population, so that the diversity of the initial population is better guaranteed.

Specifically, the input amount is the population size |p|, which has the same meaning as |p _g | in the formula (16). Firstly, searching all class 1 training task combinations meeting the total constraint of the vertical login team, namely the formula (2), and forming a class 1 training task set C ^S. Second, for the ith element C ^S (i) in C ^S, the target chromosome containing C ^S (i) is first determined according to the constraint of formula (3)Total number of tasks involved(I.e., chromosome coding length) and then based onAnd the number of tasks in C ^S (i) |C ^S (i) | construct contains C ^S (i) in lengthPermutation sets of all permutation forms in chromosomes of (a)According to the arrangement definition,The number of the elements can be calculated by the formula (20):

Third step, for Each element of (3)Firstly, taking the residual space in the corresponding chromosome as a class 2 neighborhood, and carrying out random assignment to obtain an arrangement complete codeThen calling the decoding algorithm to obtain the corresponding arrangement target valueIf it isIs a feasible solution, thenAndPooled as one individual in the initial population P ₀. Finally, the second and third steps are repeated until i= |c ^S (i) | and an initial population P ₀ including an initial code set C ₀ and an initial solution set Z ₀ is output.

In one embodiment, regarding the process of performing the crossover operation on the parent population in the step S13, the following processes may be specifically included:

Performing length transformation on two chromosome codes to be crossed, and determining crossing gene points capable of performing double-point crossing operation in the two chromosome codes to be crossed after transformation;

Two optional points in the cross gene points in one chromosome code are used as selected gene fragments, and the rest unselected gene fragments are combined with the other chromosome code to form a candidate gene library;

selecting genes and selected gene fragments from the candidate gene library to form complete codes, recovering the coding length to the original length, and decoding to obtain an initial solution;

if the initial solution is a feasible solution, ending the crossover operation and outputting new individuals in the offspring population, otherwise repeating the selection of the selected gene segments and the gene recombination processing until the initial solution is the feasible solution.

It can be understood that for the crossover operation, aiming at the problem that crossover operation is difficult to be carried out between chromosomes with different lengths, the embodiment designs a double-point crossover method based on chromosome coding length transformation, and ensures the generation of new offspring population individuals.

Specifically, the algorithm input of the crossover operation encodes the two chromosomes to be crossedAndFirst, to inputAndPerforming length transformation and determining the extended chromosome codes to be crossedAndIs a gene point capable of performing a double-point crossover operation. Subsequently, atOptionally two of the crossable gene loci of (2) as selected gene fragmentsAnd willThe remaining genes andCombining to form candidate gene libraryFinally, from the candidate gene libraryIs selected from the group consisting of a suitable gene and a selected gene fragmentComposing a complete codeRestoring it to the original lengthDecoding to obtain initial solutionIf it isIf the solution is feasible, ending the crossover operation and outputting new individuals in the child population OIf it isNot feasible, the gene fragment selection and recombination stages are repeated untilUntil feasible.

Further, in the chromosome coding length transformation stage, specifically, in the first step, the maximum common factor of the number of helicopters required by all training tasks is used for assigning value to the number of unit grouping helicopters h _min, and the extended chromosome coding length is calculated according to the assigned value

As the total number of helicopters required by all training tasks corresponding to each chromosome code in the population is always equal to the total number of helicopters which can ensure that all LHDs can move in the total movement wave times, the length of each chromosome code after expansion can be ensured to beSecond, for the input to be cross codedAndAccording to the method, each gene is coded after length transformationAndThe length of (a) is as follows:

Encoding of AndLength-transforming to obtain extended codeAndThird step, construct length andAndIdentical set of intersection front end pointsAnd a set of crossable rear end pointsFor the followingThe ith ^Full gene position in (a), if and only ifAndWhen the first coding gene is corresponding to one training task in the respective expansion codesThe other gene positions take 0 value, and the same thing applies if and only ifAndWhen the codes are the last coding gene corresponding to one training task in the respective expansion codesThe remaining gene positions were taken to be 0.

Further, in the gene fragment selection phase: first step, randomly selecting a set of crossable front pointsAnd a set of crossable rear end pointsThe gene sites i ^Front and i ^Tail with the value of 1 are required to meet the requirement of i ^Front＜i^Tial between i ^Front and i ^Tail, otherwise, the random selection operation is repeated (the existence of the condition of i ^Front＝1,i^Tail＝L^Full ensures that the condition of i ^Front＜i^Tial can be met), and the gene sites are used as the front end point and the rear end point of the selected gene fragment. Second step, willAlternatively gene fragmentsThird step, utilizeIn (a) gene composition candidate gene libraryAnd deleteAnd a gene repeated in the selected gene fragment.

Further, in the gene recombination stage: first, construct and spread the codeAndOffspring extension encoding of the same lengthThe selected gene fragmentExtension encoding as offspringFrom positions i ^Front to i ^Tail. Second, spreading and coding the offspringRemoving selected gene fragmentsAll remaining gene sites outsideAs a space to be inserted, sequentially searching candidate gene libraries for the ith ^Cross gene point in the space to be insertedSelecting the length corresponding to the single training task of all the constraints asIs used as a child extension codeFrom positions i ^Cross to i ^Cross+l^Curr -1, and deleting the corresponding coding in the candidate gene library. Thirdly, let i ^Corss←i^Corss+|σ^Curr I, repeat the second step untilUntil that point. Fourth, the newly generated child expansion codesPerforming the inverse of the chromosomal coding length transformation to obtain a normal form of the child codingDecoding to obtainFinally, ifFeasible, then output offspring individualsOtherwise, candidate gene libraryAll genes corresponding to the first training task in (a) are moved to the end, and the second to fourth steps of this paragraph are repeated (again, i ^Front =1,The presence of a condition guarantees the necessity of a loop termination).

As shown in fig. 6, a specific flow diagram of the interleaving operation of one code with another code as shown in fig. 4 is given.

In one embodiment, regarding the process of the mutation operation in the step S13, the following processes may be specifically included:

generating a random number by using 0-1 binomial distribution with probability of set mutation probability, if the random number is 1, implementing mutation, otherwise outputting the input initial chromosome code and initial solution as new individuals;

Randomly selecting two genes in the input initial chromosome codes to perform position exchange, and performing decoding operation on the chromosome codes newly generated after the exchange to obtain new solutions;

if the new solution is feasible, outputting the child individuals, otherwise, carrying out position exchange operation again until the corresponding new solution is feasible.

It will be appreciated that in this embodiment, a mutation method based on position exchange of genes in the coding is proposed, specifically, the algorithm of mutation operation is input as the initial chromosome codingInitial solutionAnd setting a mutation probability P ^Mut. In a first step, a random number B (1, P ^Mut) is generated by a 0-1 binomial distribution with probability P ^Mut, if 1, then mutation is performed, otherwiseAs a new individual output. Second step, randomly selecting input chromosome codingThe two genes in (a) are subjected to position exchange, and the newly generated chromosome is subjected to position exchangePerforming decoding operation to obtain new solutionIf it isIf not, restarting the second step untilUntil feasible. Finally, outputting the offspring individual

In one embodiment, regarding the process of the local optimizing search operation in the above step S13, the following processes may be specifically included:

after the current chromosome code is assigned as the input chromosome code, the optimal code and the optimal solution corresponding to the optimal individual are respectively assigned as the input chromosome code and the solution individual corresponding to the input chromosome code;

setting the tabu list as an empty set and calculating the number of 2 types of neighborhood genes in the input chromosome codes;

randomly selecting one 2-type neighborhood gene in the permutation complete code, randomly selecting one target sequence number for assigning a target sequence number segment of the 2-type neighborhood gene, and performing decoding operation on the updated current permutation complete code to obtain a current solution;

if the current solution is not feasible, repeating the operations of randomly selecting and arranging 2 types of neighborhood genes and decoding in the complete code until the corresponding current solution is feasible;

Judging the good-bad relation between the current solution and the optimal solution, if the pareto of the first solution is dominant, assigning 1 to the Boolean variable with the initial value of 0, and respectively updating the optimal code and the optimal solution into the complete arrangement code and the current solution;

If the Boolean variable value is 1, assigning the position of the class 2 neighborhood genes in the chromosome coding and the target sequence number segment into a tabu list;

And (3) assigning an initial value of 0 to the Boolean variable, returning to a process of randomly selecting and arranging one class 2 neighborhood gene in the complete code and randomly selecting a target sequence number segment with a target sequence number of the class 2 neighborhood gene, and performing decoding operation on the updated current complete code to obtain a current solution until the iteration times reach the algorithm iteration times, and outputting new offspring population individuals.

It will be appreciated that since the model assumption 2 above assumes a fixed number of class 2 training task consist helicopters, a dedicated helicopter fleet consist exceeding this fixed number will be encoded as a plurality of class 2 training tasks occurring adjacently to the close-range support task for a target. Therefore, unlike the class 1 training task, considering the situation of performing the group of the large machines or multi-wave grouping fire assault on a certain target, the class 2 training task in the chromosome coding is not only probably not appeared, but also probably appears many times, so that the class 2 neighborhood coding with the length of n has P ⁿ possible assignments, and the scope of the definition domain is greatly increased. In order to solve the problem, a local optimizing method based on tabu search aiming at the class 2 neighborhood is provided, and the algorithm searching efficiency is improved better.

Specifically, the input quantity of the algorithm of the local optimizing search is input chromosome codingAnd its corresponding decoded solution bodyTabu list depth D _T and algorithm iteration number N. First, the current code is assigned as valueOptimal coding corresponding to optimal individualAnd an optimal solutionRespectively assigned asAndDefining the tabu list as empty set phi, calculating input codesNumber of class 2 neighborhood genesSecond, randomly selecting and arranging the complete codesIs a class 2 neighborhood geneAnd randomly selecting a target sequence number p as the target sequence number segment of the geneAssigning a value to the updated current codePerforming decoding operation to obtain current solutionIf it isIf it is not feasible, repeating the second step untilUntil feasible.

Third, judging the current solutionAnd optimal solutionThe relationship of the merits of the first solutionPareto take advantage of (Pareto Dominate)(Denoted asIf and only ifWherein the method comprises the steps of) Then the initial value is 0 Boolean variable eta _T is assigned to be 1, and the optimal coding is carried outAnd an optimal solutionRespectively updated asAndIn addition, if no class 2 neighborhood genes have been inserted into the tabu list L _T The boolean variable η _T is also assigned a value of 1.

Fourth, if Boolean variable eta _T takes value as 1, then 2 kinds of neighborhood genes are selectedIn the specific insertion process, if the length |L _T | of the tabu list L _T reaches the assumed upper limit D _T, firstly deleting the stack bottom element of the tabu list L _T, then inserting { i, p } from the stack top, and if |L _T|＜D_T, directly inserting { i, p } into L _T. Finally, the Boolean variable eta _T is assigned an initial value of 0 again, the second step to the fourth step are repeated until the algorithm iterates for N times, and new offspring population individuals are outputIn one embodiment, the reference point is used as the basis of spatial decomposition and environmental evolution in the multi-objective evolutionary algorithm, and has a decisive influence on the optimizing performance of the multi-objective evolutionary algorithm. Taking a certain initial population (assumed that simplex coefficient h=8) generated in the foregoing as an example, firstly, the limitation of uniformly distributed reference points generated based on two traditional methods in processing population data of an amphibious helicopter training mission planning model is compared, and then a reference point self-adaptive generation method based on population distribution is provided and a calculation result example is provided.

As shown in fig. 7, the intercept between the hyperplane formed by Nadir point (lowest point) selected based on the principle of minimum euclidean distance between the hyperplane and a coordinate axis of the solution space may be smaller than 0, that is, it cannot be guaranteed that the population solution set is completely transformed into the first quadrant. In addition, because the intercept of each coordinate axis of the hyperplane and the solution space constructed by the method has randomness, the normalization processing effect on the individual target value of the solution space is not ideal.

In the prior art, a countermeasure method for the case shown in fig. 7 is also presented, that is, a hyperplane is constructed by taking the maximum value in each target dimension direction of the solution space as a Nadir point, but the method is easily affected by outliers. As shown in fig. 8, since the solution space has outliers which are significantly different from other values in the z ₁ direction, the intercept of the hyperplane in the direction is significantly increased, so that the individual solution space cannot form a relatively uniform mapping relationship with the reference point generated by the method. As shown in fig. 9, the intercept of the linear hyperplane on the z ₁ axis is significantly larger than the value of most individuals in the normalized solution set in the z ₁ axis direction, so that the individuals in the solution set are mapped on the reference point on one side of the hyperplane in a concentrated manner when clustering operation is performed according to the clustering criterion based on the interior angle measurement method, and therefore excellent individuals may be lost in subsequent selection operation of the algorithm, and the algorithm optimizing efficiency is reduced.

According to the analysis, there are two ideas that can improve the uniformity of mapping of individual solutions to reference points on hyperplane: firstly, deleting outliers with larger influence on the hyperplane intercept in the solution space, so that the hyperplane better fits most individuals in the solution set. And secondly, changing the distribution of the reference points on the hyperplane so that the reference points are gathered to the area with intensive individual deconcentration. Because the discrete coding is adopted in the method, the small difference of population individuals on the coding can cause the obvious change of the target value of the population individuals in the solution space, so the algorithm can lose the potential optimal solution by deleting the outliers in the first thought; the second idea is widely applied to some clustering-based reference point generation methods at present, but the clustering method for solving the set individuals in the methods has higher calculation complexity and is not beneficial to efficient calculation.

The self-adaptive generation method of the reference points based on population distribution can better solve the problems. Further, regarding the process of generating the reference point set on the standardized solution set by using the reference point adaptive generation method based on population distribution in the step S16, the process includes:

constructing a center vector of the standardized solution set based on median values of included angles between all individuals in the standardized solution set and each dimension direction vector;

Constructing a pseudo linear hyperplane which is perpendicular to the center vector and passes through an individual with the largest sum of all dimension target values in the standardized solution set, and obtaining the intercept of the pseudo linear hyperplane on each coordinate axis of the standardized solution space;

In each dimension direction of the standardized solution space, dividing a polar coordinate space defined by the included angle into H subspaces according to n different pitches according to the distribution condition of the included angles between an individual in the standardized solution set and the corresponding dimension direction vector; n is a partition coefficient, and H is a simplex coefficient;

Projecting boundary angles between subspaces onto coordinate axes of directions of dimension direction vectors in a standardized solution space according to the selected length to obtain H-1 separation points in a target interval on the coordinate axes;

And generating a reference point set based on the H-1 separation points on each coordinate axis of the standardized solution space.

Specifically, the input quantity is simplex coefficient H, partition coefficient n and normalized solution setFirst, a normalized solution set calculated based on a formula (17)All individuals in (3)Median value of angle theta ⁱ with each dimension direction vector e ⁱ Construction of standardized solution setsIs a center vector e _Mid of the (c). Second, construct a normalized solution set perpendicular to the center vector e _Mid The sum of all dimension target values of (a) is the largestAs shown in fig. 10 (a), to obtain the intercept of the pseudo-linear hyperplane on each coordinate axis of the normalized solution spaceThird, in each dimension direction of the normalized solution space, according to the normalized solution setIndividuals in the middleThe polar coordinate space Γi defined by the included angle theta ⁱ is divided into H subspaces according to n different distances under the distribution condition of the included angle theta ⁱ between the polar coordinate space and the dimension direction vector e ⁱ So that the subspaces are equally spacedFormed sub-spaceThe number of individual population solutions in (a) is equal, whereinFourth, subspace is formedThe boundary angle between the two is selected according to the lengthProjecting the coordinate axis in the direction of the normalized solution space e ⁱ to obtain the coordinate axisH-1 separation points in intervalFinally, based on H-1 separation points on each coordinate axis of the standardized solution spaceReferring to the prior DAS AND DENNIS (das and danish) simplex lattice method, the individual number shown in fig. 10 (b) is generated asIs provided.

If only the population ordering selection mechanism is considered, the computation complexity mainly depends on the scaling ordering method (Scalarization-based Sorting Method, SSM), the determination of Ideal point and Nadir point in the normalization process, and the segmentation operation of polar coordinate space Γ ⁱ in each dimension, where the computation complexity is O (MN ²/2), O (MN/2) and O (MN), respectively, where M is the solution space dimension, and N represents the number of individuals of the precursor population P' _g+1 (assuming 2 times the number of reference points). Therefore, the proposed reference point self-adaptive generation method based on population distribution does not increase the computational complexity in the population sorting selection stage. Furthermore, thanks to the scaled ranking method employed, the computational complexity of the population ranking selection stage is smaller than the computational complexity O (MN ²) of the non-dominant ranking based multi-objective optimization algorithm, even in the worst case.

As shown in FIG. 11, the initial population of FIG. 7 is used in the normalized spaceDirection is exemplified by individuals in a normalized solution set distributed in the first quadrantIncluded angle with e ³ The defined polar space Γ ³ is in the range of 0, pi/2. Assuming n=2, the polar coordinate spaces Γ ³ may be respectively in terms of pitchAndDivided into H subspacesSo that the synthon spaceAndThe number of solution sets of individuals in (a) is the same.

It should be noted that, although the present application uses the conventional simplex lattice method to construct the reference point, on one hand, due to the uneven segmentation of the coordinate axis, and on the other hand, due to the different intercepts of the linear hyperplane on each axis of the standard solution space, the sum of the dimension target values of the generated reference point is different and distributed on the nonlinear hypersurface.

As shown in fig. 12, the reference points in fig. 12 are distributed more densely in the region where the solution space is gathered individually, so that the solution set has better mapping effect under the condition of irregular solution set distribution, compared with the reference points uniformly distributed on the linear hyperplane in fig. 9, which is benefited by the non-uniform distribution on the curved surface. In fig. 12, each coordinate axis represents an objective function value of the planning model, and is a dimensionless number.

In some embodiments, after the case simulation is performed by using the method provided by the application, the result shows that the initial solution set generating operation and the local optimizing operation proposed by the method at the population updating iteration mechanism level fully achieve the effect of fast planning of the training task of the efficient and realistic amphibious helicopter. In practical engineering application, the application constructs an amphibious helicopter training task planning model aiming at optimization contents such as target selection, time sequence arrangement and the like of the helicopter training task, and expresses the model into a multi-target optimization model; in order to effectively evaluate the acceleration effect of helicopter training tasks on the overall situation development of landing actions, the relevant thought of system efficiency evaluation is consulted, a landing action efficiency evaluation model based on simulation is constructed, and an objective function of an amphibious helicopter training task planning model is calculated through the model; innovating in the aspects of a population updating iteration mechanism and a sorting selection mechanism respectively, the proposed algorithm has excellent performance in convergence speed and diversity of resolution, and reliable reference is provided for training task planning of the helicopter in landing actions.

It should be understood that, although the steps in the flowcharts of fig. 1 and 3 are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps of fig. 1 and 3 may include multiple sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor do the order in which the sub-steps or stages are performed necessarily occur sequentially, but may be performed alternately or alternately with at least some of the sub-steps or stages of other steps or steps.

As shown in fig. 13, in one embodiment, there is further provided an adaptive multi-objective task planning system 100 based on reference points, including a parameter acquisition module 11, a model calling module 12, a population construction module 13, a child generation module 14, a merging calculation module 15, a set generation module 16, a new population module 17, and an iteration control module 18. Wherein:

The parameter acquisition module 11 is used for acquiring known task parameters; the task parameters include the total number of platforms, the total wave number of helicopter play, the solution space dimension, the simplex segmentation parameters and the maximum iteration number. The model calling module 12 is used for calling the constructed amphibious helicopter training task planning model; constraint conditions of the amphibious helicopter training task planning model comprise vertical landing team quantity constraint, helicopter quantity constraint, platform guarantee resource constraint, task stage time sequence constraint based on execution sequence and task stage time sequence constraint based on platform deck scheduling, and optimization objective functions of the amphibious helicopter training task planning model comprise time-consuming objective functions of a minimum landing stage, a minimum manpower loss objective function and a minimum ground threat objective function of the helicopter.

The population construction module 13 is used for calculating the population scale of the amphibious helicopter training mission planning model based on the mission parameters and constructing an initial population. The offspring generation module 14 is configured to perform crossover operations, mutation operations, and local optimization search operations on the parent population to generate offspring populations. The merging calculation module 15 is configured to merge the generated child population and the parent population to obtain a merged population, update the IDEAL point and the NADIR point according to the elements in the solution set corresponding to the merged population, and calculate the normalized solution set according to the IDEAL point and the NADIR point. The set generating module 16 is configured to generate a set of reference points and a set of corresponding interest solutions on the standardized solution set using a population distribution-based reference point adaptive generation method. The new population module 17 is configured to assign each individual in the normalized solution set to the interest solution set by using a clustering method based on an interior angle measurement method, perform a non-dominant ranking operation in each interest solution set based on a scaled ranking method, and select individuals corresponding to the top-ranked elements to form a new generation population. The iteration control module 18 is used for outputting amphibious helicopter training task planning scheme data when the iteration number of the population reaches the maximum iteration number.

The adaptive multi-objective mission planning system 100 based on the reference point provides an amphibious helicopter training mission planning model based on reasonable assumption of a real landing training scene by refining and analyzing main mission modes and characteristics of the helicopter in the amphibious landing action, and expresses the model as a multi-objective optimization model, and aims to minimize the time consumption of landing actions, casualties of a host side and the air fire threat of the host side of the helicopter under blue side. The amphibious helicopter training task planning model considers the specific reality constraint of two aspects of training resources which can be used by multiple platforms, training and guaranteeing the full-link task stage time sequence. The adaptive generation method of the reference points based on the population distribution is applied to the population sorting selection mechanism level, the optimizing capability of the irregular population is enhanced more effectively on the premise of not increasing the computational complexity, new initial solution generation, intersection, variation and local optimizing methods are provided to the population updating iteration mechanism level, the searching efficiency of the coding definition space is improved better, good performance is achieved in optimizing efficiency and maintaining population diversity, and therefore the purpose of fast planning of the training task of the amphibious helicopter is achieved more efficiently and effectively.

In one embodiment, the modules of the adaptive multi-objective mission planning system 100 based on the reference point may be further used to implement further processing steps in the embodiments of the adaptive multi-objective mission planning method based on the reference point.

For a specific definition of the adaptive multi-objective mission planning system 100 based on the reference point, reference may be made to the corresponding definition of the adaptive multi-objective mission planning method based on the reference point hereinabove, and will not be described herein. The various modules in the adaptive multi-objective mission planning system 100 described above based on reference points may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be stored in a memory of the above device, or may be stored in software, so that the processor may call and execute operations corresponding to the above modules, where the above device may be, but is not limited to, various computer devices existing in the art.

In one embodiment, there is also provided a computer device including a memory and a processor, the memory storing a computer program, the processor executing the computer program to perform the steps of: acquiring known task parameters; the task parameters comprise the total number of platforms, the total wave number of helicopter running, the dimension of solution space, the simplex segmentation parameter and the maximum iteration number; calling a constructed amphibious helicopter training task planning model; constraint conditions of the amphibious helicopter training task planning model comprise vertical login team quantity constraint, helicopter quantity constraint, platform guarantee resource constraint, task stage time sequence constraint based on execution sequence and task stage time sequence constraint based on platform deck scheduling, and optimization objective functions of the amphibious helicopter training task planning model comprise time-consuming objective functions of a minimum login stage, a minimum manpower loss objective function and a minimum ground threat objective function of the helicopter;

Calculating a population scale for an amphibious helicopter training task planning model based on task parameters and constructing an initial population; performing crossover operation, mutation operation and local optimization search operation on the parent population to generate a child population; combining the generated child population with the parent population to obtain a combined population, updating IDEAL points and NADIR points according to elements in a solution set corresponding to the combined population, and calculating according to the IDEAL points and NADIR points to obtain a standardized solution set; generating a reference point set and a corresponding set of interest solution sets on the standardized solution set by using a reference point self-adaptive generation method based on population distribution; utilizing a clustering method based on an interior angle measurement method to distribute each individual in the standardized solution set to the interest base solution set, carrying out non-dominant sorting operation in each interest base solution set based on a scaled sorting method, and selecting individuals corresponding to elements ranked at the front to form a new generation population; and returning to the step of executing the cross operation, the mutation operation and the local optimizing search operation on the parent population to generate the child population, and outputting amphibious helicopter training task planning scheme data when the iteration number of the population reaches the maximum iteration number.

It will be appreciated that the above-mentioned computer device may include other software and hardware components not listed in the specification besides the above-mentioned memory and processor, and may be specifically determined according to the model of the specific computer device in different application scenarios, and the detailed description will not be listed in any way.

In one embodiment, the processor may further implement the steps or sub-steps added to the embodiments of the adaptive multi-objective task planning method based on the reference point when executing the computer program.

In one embodiment, there is also provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of: acquiring known task parameters; the task parameters comprise the total number of platforms, the total wave number of helicopter running, the dimension of solution space, the simplex segmentation parameter and the maximum iteration number; calling a constructed amphibious helicopter training task planning model; constraint conditions of the amphibious helicopter training task planning model comprise vertical login team quantity constraint, helicopter quantity constraint, platform guarantee resource constraint, task stage time sequence constraint based on execution sequence and task stage time sequence constraint based on platform deck scheduling, and optimization objective functions of the amphibious helicopter training task planning model comprise time-consuming objective functions of a minimum login stage, a minimum manpower loss objective function and a minimum ground threat objective function of the helicopter;

Calculating a population scale for an amphibious helicopter training task planning model based on task parameters and constructing an initial population; performing crossover operation, mutation operation and local optimization search operation on the parent population to generate a child population; combining the generated child population with the parent population to obtain a combined population, updating IDEAL points and NADIR points according to elements in a solution set corresponding to the combined population, and calculating according to the IDEAL points and NADIR points to obtain a standardized solution set; generating a reference point set and a corresponding set of interest solution sets on the standardized solution set by using a reference point self-adaptive generation method based on population distribution; utilizing a clustering method based on an interior angle measurement method to distribute each individual in the standardized solution set to the interest base solution set, carrying out non-dominant sorting operation in each interest base solution set based on a scaled sorting method, and selecting individuals corresponding to elements ranked at the front to form a new generation population; and returning to the step of executing the cross operation, the mutation operation and the local optimizing search operation on the parent population to generate the child population, and outputting amphibious helicopter training task planning scheme data when the iteration number of the population reaches the maximum iteration number.

In one embodiment, the computer program may further implement the steps or sub-steps added in the embodiments of the adaptive multi-objective task planning method based on the reference point.

Those skilled in the art will appreciate that implementing all or part of the above-described methods in accordance with the embodiments may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus dynamic random access memory (Rambus DRAM, RDRAM for short), and interface dynamic random access memory (DRDRAM), among others.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The foregoing examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it is possible for those skilled in the art to make several variations and modifications without departing from the spirit of the present application, which fall within the protection scope of the present application. The scope of the application is therefore intended to be covered by the appended claims.

Claims (11)

1. The adaptive multi-target task planning method based on the reference point is characterized by comprising the following steps:

Acquiring known task parameters; the task parameters comprise the total number of platforms, the total wave number of helicopter running, the solution space dimension, the simplex segmentation parameter and the maximum iteration number;

Calling a constructed amphibious helicopter training task planning model; constraint conditions of the amphibious helicopter training task planning model comprise vertical landing team quantity constraint, helicopter quantity constraint, platform guarantee resource constraint, task stage time sequence constraint based on execution sequence and task stage time sequence constraint based on platform deck scheduling, and optimization objective functions of the amphibious helicopter training task planning model comprise a time-consuming objective function of a minimum landing stage, a labor-wasting objective function and a ground threat objective function of a minimum helicopter;

Calculating a population scale for the amphibious helicopter training task planning model based on the task parameters and constructing an initial population;

Performing crossover operation, mutation operation and local optimization search operation on the parent population to generate a child population;

Combining the generated offspring population with the parent population to obtain a combined population, updating IDEAL points and NADIR points according to elements in a solution set corresponding to the combined population, and calculating according to the IDEAL points and the NADIR points to obtain a standardized solution set;

Generating a reference point set and a corresponding set of interest solution sets on the standardized solution set by using a reference point self-adaptive generation method based on population distribution;

utilizing a clustering method based on an interior angle measurement method to distribute each individual in the standardized solution set to the interest solution set, performing non-dominant sorting operation in each interest solution set based on a scaled sorting method, and selecting individuals corresponding to elements ranked at the front to form a new generation population;

And returning to the step of executing the cross operation, the mutation operation and the local optimizing search operation on the parent population to generate the child population, and outputting amphibious helicopter training task planning scheme data until the iteration number of the population reaches the maximum iteration number.

2. The adaptive multi-objective mission planning method based on reference points as claimed in claim 1, wherein the chromosome coding method adopted in the population iteration process is as follows:

Decomposing all training tasks meeting the condition of setting the number of helicopter groups into a single task list with the number of helicopter groups equal to 1 and adopting discrete double-row chromosome coding;

adding a vacancy gene in the chromosome coding of the task list;

The chromosome decoding mode adopted in the population iteration process is as follows:

converting the chromosome codes of the task list into a play list representing the play plan of each helicopter;

Obtaining an initiating time list of a vertical login task and a close-range support task for each target through the play list;

According to the play list, the start time list and the scheduled training plan of the plane login team, the change of the weapon force of each target node of the login field along with time is calculated by utilizing an improved Lanchester model, a login field node state vector is obtained, the progress direction of login actions is deduced through a login field situation transfer model, and the achievement condition of the training purpose and the training task efficiency index set are determined.

3. The adaptive multi-objective mission planning method of claim 2, wherein the converting to the play list includes:

calculating the total number of vertical login teams delivered by each vertical login task in the chromosome coding of the task list and the number of various helicopters required by each vertical login task and a close-range support task;

selecting 1 class of training tasks meeting training resource constraint in the task list according to the total vertical login team quantity and the various helicopter quantity, assigning values to elements in the play list according to a set assignment formula, and adding the executed tasks to the executed task list according to the original list sequence; the class 1 training task is the vertical login task;

After all 1-class training tasks meeting the training resource constraint are inserted into the play list, selecting 2-class training tasks meeting the training resource constraint in the task list according to the total vertical login team quantity and the various helicopter quantity, and correspondingly updating the play list and the executed task list; the class 2 training task is the close-range support task;

The process for solving the starting time list comprises the following steps:

Drawing a minimum risk route set from the helicopter to each task target by using Dijiestra algorithm according to the task targets of each training task in the play list and the distribution condition of target air defense nodes in a landing field;

solving the starting time list of the task execution stage of each training task by utilizing the starting list and the route set;

the process for determining the training task performance index set comprises the following steps:

setting the state vector value of the corresponding landing field node at the moment of initiating the landing action to be zero;

solving the bearing risk in the flight process of each helicopter grouping task based on the setting;

According to the starting time list and the scheduled training plan of the plane login team, finding out the login field node which is in antagonism in each simulation time, calculating the manpower loss and time consumption of each login field node against both sides, updating the corresponding login field node state vector according to the login field situation transfer condition, and carrying out numerical integration operation on the manpower force vector of both sides at each target node by setting the simulation step length until the login field node state vector meets the scheduled target of the login action during the login field opening stage or the training target cannot be achieved;

And counting total consumed time at the open-up landing stage, accumulated manpower loss of the vertical landing team and the plane landing team and bearing risks of all helicopters executing the training tasks to form the training task efficiency index set.

4. A method of adaptive multi-objective mission planning based on reference points as claimed in any one of claims 1 to 3, wherein the process of constructing the initial population includes:

searching all 1-class training task combinations meeting the vertical login team quantity constraint to form a 1-class training task set;

determining the total task number contained in the target chromosome according to the helicopter quantity constraint; the target chromosome is a chromosome containing elements in the class 1 training task set;

Constructing an arrangement set containing all arrangement forms of the corresponding elements in the class 1 training task set in a chromosome with the length of the total task number based on the total task number and the task number contained in the corresponding elements in the class 1 training task set;

For each element in the arrangement set, taking the residual space in the chromosome corresponding to the element as a class 2 neighborhood and carrying out random assignment to obtain an arrangement complete code; the class 2 neighborhood is a neighborhood in which a class 2 training task corresponding gene is located;

chromosome decoding is carried out on the permutation complete codes to obtain corresponding permutation target values;

If the ranking target value is a feasible solution, combining the ranking complete code and the corresponding ranking target value as an individual in the initial population;

And returning to the step of determining the total number of tasks contained in the target chromosome according to the helicopter number constraint until each task contained in each element in the permutation collection is traversed, and outputting the initial population containing an initial coding collection and an initial solution collection.

5. A method of adaptive multi-objective mission planning based on reference points as claimed in any one of claims 1 to 3, in which the process of performing crossover operations on parent populations includes:

Performing length transformation on two chromosome codes to be crossed, and determining crossing gene points capable of performing double-point crossing operation in the two chromosome codes to be crossed after transformation;

two optional points in the cross gene points in one chromosome code are taken as selected gene fragments, and the rest unselected gene fragments are combined with the other chromosome code to form a candidate gene library;

selecting genes and the selected gene fragments from the candidate gene library to form complete codes, recovering the coding length to the original length, and decoding to obtain an initial solution;

And if the initial solution is a feasible solution, ending the crossover operation and outputting new individuals in the offspring population, otherwise, repeating the selection of the selected gene segments and the gene recombination treatment until the initial solution is the feasible solution.

6. A method of adaptive multi-objective mission planning based on a reference point as claimed in any one of claims 1 to 3, in which the process of mutation operation includes:

Generating a random number through 0-1 binomial distribution with probability of set mutation probability, if the random number is 1, implementing mutation, otherwise outputting the input initial chromosome code and initial solution as new individuals;

Randomly selecting two genes in the input initial chromosome codes to exchange positions, and decoding the chromosome codes newly generated after exchange to obtain new solutions;

And if the new solution is feasible, outputting child individuals, otherwise, carrying out position exchange operation again until the new solution is feasible correspondingly.

7. A method of adaptive multi-objective mission planning based on reference points as claimed in any one of claims 1 to 3, in which the process of local optimization search operation includes:

assigning the current chromosome code as an input chromosome code, and respectively assigning the optimal code and the optimal solution corresponding to the optimal individual as an input chromosome code and a solution individual corresponding to the input chromosome code;

Setting a tabu list as an empty set and calculating the number of 2 types of neighborhood genes in the input chromosome codes;

Randomly selecting one 2-type neighborhood gene in the permutation complete code, randomly selecting one target sequence number for assigning a target sequence number segment of the 2-type neighborhood gene, and performing decoding operation on the updated current permutation complete code to obtain a current solution;

If the current solution is not feasible, repeating the operations of randomly selecting and arranging 2 types of neighborhood genes and decoding in the complete code until the corresponding current solution is feasible;

Judging the good-bad relation between the current solution and the optimal solution, if the pareto of the first solution is dominant, assigning a Boolean variable with an initial value of 0 to be 1, and respectively updating the optimal code and the optimal solution into the arranged complete code and the current solution;

If the Boolean variable value is 1, assigning the position of the class 2 neighborhood genes in the chromosome coding and the target sequence number segment into the tabu list;

And (3) assigning an initial value of 0 to the Boolean variable, returning to a process of randomly selecting one 2-class neighborhood gene in the permutation complete code and randomly selecting one target sequence number to assign a target sequence number segment of the 2-class neighborhood gene, and performing decoding operation on the updated permutation complete code to obtain a current solution until the iteration times reach the algorithm iteration times, and outputting new child population individuals.

8. The reference point-based adaptive multi-objective mission planning method of claim 1, wherein generating a set of reference points on the standardized solution set using a population distribution-based reference point adaptive generation method comprises:

constructing a center vector of the standardized solution set based on median values of included angles of all individuals in the standardized solution set and each dimension direction vector;

Constructing a pseudo linear hyperplane which is perpendicular to the center vector and passes through the maximum sum of all dimension target values in the standardized solution set, and obtaining the intercept of the pseudo linear hyperplane on each coordinate axis of the standardized solution space;

In each dimension direction of a standardized solution space, dividing a polar coordinate space defined by the included angle into H subspaces according to n different pitches according to the distribution condition of the included angles between an individual in the standardized solution set and the corresponding dimension direction vector; n is a partition coefficient, and H is a simplex coefficient;

Projecting boundary angles between subspaces onto coordinate axes of directions of dimension direction vectors in a standardized solution space according to the selected length to obtain H-1 separation points in a target interval on the coordinate axes;

and generating the reference point set based on H-1 separation points on each coordinate axis of the standardized solution space.

9. An adaptive multi-objective mission planning system based on a reference point, comprising:

the parameter acquisition module is used for acquiring known task parameters; the task parameters comprise the total number of platforms, the total wave number of helicopter running, the solution space dimension, the simplex segmentation parameter and the maximum iteration number;

The model calling module is used for calling the constructed amphibious helicopter training task planning model; constraint conditions of the amphibious helicopter training task planning model comprise vertical landing team quantity constraint, helicopter quantity constraint, platform guarantee resource constraint, task stage time sequence constraint based on execution sequence and task stage time sequence constraint based on platform deck scheduling, and optimization objective functions of the amphibious helicopter training task planning model comprise a time-consuming objective function of a minimum landing stage, a labor-wasting objective function and a ground threat objective function of a minimum helicopter;

The population construction module is used for calculating the population scale of the amphibious helicopter training task planning model based on the task parameters and constructing an initial population;

the offspring generation module is used for executing crossover operation, mutation operation and local optimization search operation on the parent population to generate offspring population;

the merging calculation module is used for merging the generated child population with the parent population to obtain a merged population, updating IDEAL points and NADIR points according to elements in a solution set corresponding to the merged population, and calculating according to the IDEAL points and the NADIR points to obtain a standardized solution set;

the set generation module is used for generating a reference point set and a set of corresponding interest solution sets on the standardized solution set by using a reference point self-adaptive generation method based on population distribution;

A new population module, configured to assign each individual in the standardized solution set to the interest solution set by using a clustering method based on an interior angle measurement method, perform a non-dominant ranking operation in each of the interest solution sets based on a scaled ranking method, and select individuals corresponding to the top-ranked elements to form a new generation population;

And the iteration control module is used for outputting amphibious helicopter training task planning scheme data when the population iteration number reaches the maximum iteration number.

10. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the reference point based adaptive multi-objective task planning method according to any one of claims 1 to 7 when the computer program is executed.

11. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the reference point based adaptive multi-objective task planning method according to any one of claims 1 to 7.