CN106371318A - Facility environment multi-objective optimization control method based on cooperative game - Google Patents

Abstract

The invention provides a facility environment multi-objective optimization control method based on cooperative game. The method comprises the steps that S1 three control objectives are established for a facility environment, wherein an environmental parameter-temperature and humidity control objective, an energy control objective and control device physical limit, namely a device loss control target, are comprised; S2, a cooperative game model is established, and specifically a multi-objective model predictive control method based on the cooperative game model is used to process three control objectives of the facility environment to acquire the optimal function of each control objective; and S3, the optimal function of the objectives is solved to acquire the optimal controlled quantity of each objective function. According to the invention, the game theory idea is used; each control objective is taken as a game side; through the establishment of the cooperative competition model, each game side takes into account the individual interests and the interests of others at the same time; and a multi-objective conflict problem is solved.

Description

Cooperative game-based facility environment multi-objective optimization control method

Technical Field

The invention relates to the field of intelligent agriculture, in particular to a cooperative game-based facility environment multi-objective optimization control method.

Background

At present, the modern agricultural production mode, the integrated application perception identification technology, the monitoring technology, the communication network technology, the cloud computing and big data technology and other ubiquitous network (internet and internet of things) technologies are taken as remarkable characteristics, so that the intelligent control, the intelligent production and the whole-process quality control and traceability of agricultural products in the agricultural production process are realized, and the intelligent, industrialized, standardized, efficient and intensive green and low-carbon modern agricultural industry embodied by agricultural 4.0 is provided. The facility agriculture is a symbol of the development of modern agriculture, is one of the most active industries in the world at present, is a main technical measure for providing fresh agricultural products in all countries of the world, and the intelligent (environmental) control advocated by the agriculture 4.0 is a key technology of the facility agriculture.

Generally, the main control methods in a facility environment are pid (proportion integration differentiation) control and fuzzy control, and in recent years, due to the development of Model Predictive Control (MPC) technology and the huge potential of processing complex constraint optimization control problems appearing in industrial processes, the application of MPC is extended from traditional industrial processes, such as oil refining, petrochemical industry and chemical industry, to the fields of electric power, steel, ships, environment, medical treatment, agriculture and the like. The traditional greenhouse control only considers the control of a single target, namely, factors such as temperature, humidity, illumination and nutrition of a facility environment are required to reach a proper value, and other factors such as energy consumption and other control targets such as physical limitation of equipment are not considered, so that a theoretical optimization control algorithm cannot be effectively applied to an actual facility environment due to high operation cost. Therefore, in the control of facility environment, it is important to consider the energy consumption of production and the loss of equipment. The control strategy not only can ensure the optimal environment required by the growth of crops so as to maximize the crop yield, but also can effectively reduce the energy consumption and the production cost of greenhouse production so as to obtain greater economic benefit, and is receiving more and more attention. It follows that control of the facility environment is in fact a multi-objective control problem.

In multi-objective optimization designs, the objectives generally conflict with each other, for example, in plant environment control, between precise control and control of energy consumption. To reconcile these conflicts, some way of resolving the conflict between the objectives is required. The game theory is a problem of how to make decisions and how to balance the decisions when behaviors of decision subjects interact directly, and the game theory is applied more and more in the field of engineering design at present. In view of the similarity between the multi-objective optimization problem and the game problem, the game theory idea and method can be introduced into the solution of the multi-objective optimization design problem so as to overcome the defects of the traditional multi-objective optimization design problem solution method.

Disclosure of Invention

The invention aims to provide a cooperative game-based facility environment multi-objective optimization control method aiming at the problem of multi-objective control in a facility environment. And decomposing the whole control system into a plurality of subsystems according to the control target, and obtaining the whole optimum of the system by optimizing respective objective functions of each subsystem. In order to realize balanced solution of each control target, each control target is regarded as each game party by using the thought of game theory for reference, and each game party considers the interests of others while considering the interests of individuals by establishing a cooperative competition model, so that the problem of multi-target conflict is solved.

The technical scheme of the invention is as follows:

a cooperative game-based facility environment multi-objective optimization control method comprises the following steps:

s1, three control targets of the facility environment are established: environmental parameters-temperature and humidity control target, energy consumption control target and control equipment physical limit, namely equipment loss control target;

s2, establishing a cooperative game model, and processing three control targets of the facility environment by adopting a multi-target model prediction control method based on the cooperative game model to obtain an optimization function of each control target;

and S3, solving the obtained optimization function of the target, and obtaining the optimal control quantity of each target function.

Step S1 of the present invention specifically includes:

s1-1, establishing an environment parameter-temperature and humidity control objective function f of the facility₁The description is as follows:

$f_1=min\[\underset{j=1}{\overset{p}{\Σ}}{\left(T_{in}\right(k+j|k)-T_{set})}^2+\underset{j=1}{\overset{p}{\Σ}}{\left(H_{in}\right(k+j|k)-H_{set})}^2\]$

wherein: p represents a prediction time domain, j represents a prediction step length, and k represents the current time; t is_in(k + j | k) represents the indoor temperature (obtained from the facility greenhouse microclimate prediction model) at the future time k + j obtained at the current time k, H_in(k + j | k) represents the indoor humidity (obtained from the facility greenhouse microclimate prediction model) at the future time k + j obtained at the current time k, T_setAnd H_setRespectively the temperature value and the humidity value which are suitable for the growth of crops;

s1-2, establishing an energy consumption control objective function f for facility control₂The description is as follows:

$f_2=min\[\underset{i=1}{\overset{N}{\Σ}}\underset{j^{\′}=0}{\overset{M-1}{\Σ}}\left(u_i\right(k+j^{\′}|k)/u_{i\_max})/M\]$

wherein N represents the total number of control devices, and i represents the number of the current control device; m represents a control time domain, and j' represents a control step size; u. of_i(k + j' | k) represents the optimal control quantity (obtained according to the optimal objective function of the control subsystem) of the future k + j moment control equipment i acquired at the current moment k, and u_{i_max}A maximum control amount for the control device i (maximum power of the control device such as a heater in a facility, maximum opening angle of a louver, number of fans to be opened, etc.);

s1-3, establishing a device loss control objective function f in the facility₃The description is as follows:

$f_3=min\underset{i=1}{\overset{N}{\Σ}}\left(u_i\right(k+j^{\′}+1|k)-u_i(k+j^{\′}|k\left)\right)/u_{i\_\max }$

wherein: u. of_iAnd (k + j' +1| k) represents the optimal control quantity (obtained according to the optimal objective function of the control subsystem) of the control equipment i at the future time k + j +1 acquired at the current time k.

Step S2 of the present invention specifically includes:

s2-1, respectively taking three control targets, namely three facility control subsystems, as game parties, and establishing a cooperative game model by adopting the following formula:

$s_l\left(u_l^{*}\right)=\underset{u_l\∈U_l,u_{l^{\′}}\∈U_{l^{\′}}}{max}\[w_{ll}\×{\stackrel{\‾}{s}}_l(u_l,u_{l^{\′}})+\underset{l^{\′}=1(l^{\′}\≠l)}{\overset{m}{\Σ}}{w}_{{\mathrm{ll}}^{\′}}\×{\hat{s}}_{{\mathrm{ll}}^{\′}}(u_l,u_{l^{\′}})\]$

where l and l' denote the number of the gambling parties and m denotes the total number of gambling parties; u shape_lAnd U_l′Set of policies, u, representing gambling parties l and l_lAnd u_l′Indicating the strategy employed by the gambling parties l and l',represents the optimal strategy that the gambling party l can adopt;representing the adoption of policy u by gambling parties l and l', respectively_lAnd u_l′The absolute benefit of the gambling party l,representing the adoption of policy u by gambling parties l and l', respectively_lAnd u_l′The absolute benefit of the betting party l' in time,indicating that betting party l takes an optimal strategyThe maximum absolute income sum which can be obtained by all game parties; w is a_llAnd w_ll′Is a proportionality coefficient, w_llIndicating the extent of non-malfunction of the gambling party l, w_ll′Indicates the degree of cooperation of the gambling party l, and

s2-2, establishing a multi-target model predictive control method based on the cooperative game model, and obtaining an optimized target function of each subsystem in the facility environment as follows:

$S_l\left(u_{l\_i}^{*}\right)=\max \[w_{ll}\×\frac{f_l(u_{l\_i}^{*},u_{l^{\′}\_i^{\′}})}{f_l(u_{l\_i},u_{l^{\′}\_i^{\′}})}+\underset{l^{\′}=1(l^{\′}\≠l)}{\overset{3}{\Σ}}{w}_{{\mathrm{ll}}^{\′}}\×\frac{f_{l^{\′}}(u_{l\_i}^{*},u_{l^{\′}\_i^{\′}})}{f_{l^{\′}}(u_{l\_i},u_{l^{\′}\_i^{\′}})}\]$

wherein: l and l' denote the number of the facility control subsystem, i.e. the gaming party; u. of_{l_i}And u_{l′_i′}The control quantity of the control devices i and i 'in the control subsystems l and l', namely the strategy adopted by the game party, is represented;indicating the control subsystem lThe optimal control quantity of the internal control equipment i, namely an optimal strategy; f. of_lAnd f_l′Representing the control objective function, i.e. the absolute gain, of the control subsystems l and l'; f. of_l(u_{l_i},u_{l′_i′}) And f_l′(u_{l_i},u_{l′_i′}) Denotes a control quantity u of the control devices i and i' in the control subsystems l and l_l，iAnd u_l′，i′Absolute gain of each subsystem;which represents the maximum absolute gain of the entire control system when the control device i in the control subsystem l employs the optimum control quantity.

Step S3 of the present invention specifically includes:

s3-1, respectively randomly initializing control sequences u of respective control devices by each control subsystem_{l_i}(0)；

u_{l_i}(0)＝[u_{l_i}(0|0),u_{l_i}(1|0),...,u_{l_i}(M-1|0)],l＝1,2,3

Wherein: m denotes the control time domain, l denotes the control subsystem number, i denotes the number of the control device, u_{l_i}(0|0) represents a random preset value of the control quantity at the 0 th moment of the control equipment i in the control subsystem l, u_{l_i}(M-1|0) represents a random preset value of the control quantity of the control equipment i at the moment 0 in the control subsystem I at the moment M-1;

s3-2, the control subsystems communicate with each other and respectively control the control sequence u of the time k_{l_i,q}(k) Sending the data to other subsystems;

u_{l_i,q}(k)＝[u_{l_i}(k|k),u_{l_i}(k+1|k),...,u_{l_i}(k+M-1|k)]^q

wherein: u. of_{l_i}(k + M-1| k) represents a preset value of a control quantity of a control device i at the moment k to the control quantity of the control device i at the moment k + M-1 in a control subsystem l, and q represents the number of times of the control quantity of the control subsystem l and other subsystems at the moment k;

s3-3, after obtaining the control sequence sent by other control subsystems for the q times at the moment k, the control subsystem l obtains the optimal control sequence by the particle swarm optimization according to the own optimization objective function

$u_{l\_i,q}^{*}\left(k\right)={\[u_{l\_i}^{*}\left(k\right|k),u_{l\_i}^{*}(k+1|k),...,u_{l\_i}^{*}(k+M-1|k)\]}^q$

Wherein,the optimal preset value of the control quantity of the control equipment i at the time k to the time k + M-1 in the control subsystem l is represented;

s3-4, for any one of the gambling parties, if q is satisfied at time k, q is equal to q_maxOrThe transmission of the control sequence is stopped and the optimal control sequence is executed1 st control quantity ofThereafter, the process returns to step S4-2 to calculate the optimal control sequence at time k + 1.

Wherein q is_maxA threshold value representing the maximum number of times the control sequence is sent by the control subsystem at a time, representing the improvement in the control sequence.

The invention has the beneficial effects that:

aiming at the multi-target control problem in the facility environment, the distributed model predictive control method is adopted, the overall control system is decomposed into a plurality of subsystems according to the control target, and each subsystem obtains the overall optimization of the system by optimizing the respective target function. In order to realize balanced solution of each control target, each control target is regarded as each game party by using the thought of game theory for reference, and each game party considers the interests of others while considering the interests of individuals by establishing a cooperative competition model, so that the problem of multi-target conflict is solved. For convenience of explanation, the effectiveness of the algorithm is verified on an MATLAB simulation platform by taking a multi-target control system consisting of temperature and humidity parameters of a facility environment and corresponding control equipment as an example (the multi-target control system can be expanded to a multi-parameter forming multi-target control system).

The invention researches the multi-target control problem in the facility environment, provides an application framework of the distributed model predictive control in the facility environment aiming at the nonlinearity of the facility environment model and the complexity and the constraint of a target function, and solves the conflict problems of the three targets of accurate control, minimum energy consumption and minimum equipment opening frequency in the facility environment by adopting a cooperative game method by using the game theory for reference. Simulation and verification are carried out on an MATLAB platform by taking meteorological data of spring, summer, autumn and winter of 2015 as an example, and the meteorological data are compared with single-target control and traditional linear weighted multi-target control. Experimental results show that the algorithm provided by the invention has obvious advantages in solving the problem of multi-target conflict, the indoor temperature and humidity can meet the appropriate range required by crop growth, and meanwhile, the energy consumption of control equipment and the equipment loss are greatly reduced.

Detailed Description

The present invention will be further described with reference to the following examples.

A cooperative game-based facility environment multi-objective optimization control method comprises the following steps:

s1, three control targets of the facility environment are established: environmental parameters-temperature and humidity control target, energy consumption control target and control equipment physical limit, namely equipment loss control target;

s1-1, establishing an environment parameter-temperature and humidity control objective function f of the facility₁The description is as follows:

$f_1=min\[\underset{j=1}{\overset{P}{\Σ}}{\left(T_{in}\right(k+j|k)-T_{set})}^2+\underset{j=1}{\overset{P}{\Σ}}{\left(H_{in}\right(k+j|k)-H_{set})}^2\]$

wherein: p represents a prediction time domain, j represents a prediction step length, and k represents the current time; t is_in(k + j | k) represents the indoor temperature (obtained from the facility greenhouse microclimate prediction model) at the future time k + j obtained at the current time k, H_in(k + j | k) represents the indoor humidity (obtained from the facility greenhouse microclimate prediction model) at the future time k + j obtained at the current time k, T_setAnd H_setRespectively the temperature value and the humidity value which are suitable for the growth of crops;

s1-2, establishing an energy consumption control objective function f for facility control₂DrawingThe method comprises the following steps:

$f_2=min\[\underset{i=1}{\overset{N}{\Σ}}\underset{j=0}{\overset{M-1}{\Σ}}\left(u_i\right(k+j^{\′}|k)/u_{i\_max})/M\]$

wherein N represents the total number of control devices, and i represents the number of the current control device; m represents a control time domain, and j' represents a control step size; u. of_i(k + j' | k) represents the optimal control quantity (obtained according to the optimal objective function of the control subsystem) of the future k + j moment control equipment i acquired at the current moment k, and u_{i_max}A maximum control amount for the control device i (maximum power of the control device such as a heater in a facility, maximum opening angle of a louver, number of fans to be opened, etc.);

s1-3, establishing a device loss control objective function f in the facility₃The description is as follows:

$f_3=min\underset{i=1}{\overset{N}{\Σ}}\left(u_i\right(k+j^{\′}+1|k)-u_i(k+j^{\′}|k\left)\right)/u_{i\_max}$

wherein: u. of_iAnd (k + j' +1| k) represents the optimal control quantity (obtained according to the optimal objective function of the control subsystem) of the control equipment i at the future time k + j +1 acquired at the current time k.

S2, establishing a cooperative game model, and processing three control targets of the facility environment by adopting a multi-target model prediction control method based on the cooperative game model to obtain an optimization function of each control target;

s2-1, respectively taking three control targets, namely three facility control subsystems, as game parties, and establishing a cooperative game model by adopting the following formula:

$s_l\left(u_l^{*}\right)=\underset{u_l\∈U_l,u_{l^{\′}}\∈U_{l^{\′}}}{max}\[w_{ll}\×{\stackrel{\‾}{s}}_l(u_l,u_{l^{\′}})+\underset{l^{\′}=1(l^{\′}\≠l)}{\overset{m}{\Σ}}{w}_{{\mathrm{ll}}^{\′}}\×{\hat{s}}_{{\mathrm{ll}}^{\′}}(u_l,u_{l^{\′}})\]$

where l and l' denote the number of the gambling parties and m denotes the total number of gambling parties; u shape_lAnd U_l′Set of policies, u, representing gambling parties l and l_lAnd u_l′Indicating the strategy employed by the gambling parties l and l',represents the optimal strategy that the gambling party l can adopt;representing the adoption of policy u by gambling parties l and l', respectively_lAnd u_l′The absolute benefit of the gambling party l,representing the adoption of policy u by gambling parties l and l', respectively_lAnd u_l′The absolute benefit of the betting party l' in time,indicating that betting party l takes an optimal strategyThe maximum absolute income sum which can be obtained by all game parties; w is a_llAnd w_ll′Is a proportionality coefficient, w_llIndicating the extent of non-malfunction of the gambling party l, w_ll′Indicates the degree of cooperation of the gambling party l, and

s2-2, establishing a multi-target model predictive control method based on the cooperative game model, and obtaining an optimized target function of each subsystem in the facility environment as follows:

$S_l\left(u_{l\_i}^{*}\right)=\max \[w_{ll}\×\frac{f_l(u_{l\_i}^{*},u_{l^{\′}\_i^{\′}})}{f_l(u_{l\_i},u_{l^{\′}\_i^{\′}})}+\underset{l^{\′}=1(l^{\′}\≠l)}{\overset{3}{\Σ}}{w}_{{\mathrm{ll}}^{\′}}\×\frac{f_{l^{\′}}(u_{l\_i}^{*},u_{l^{\′}\_i^{\′}})}{f_{l^{\′}}(u_{l\_i},u_{l^{\′}\_i^{\′}})}\]$

wherein: l and l' denote the number of the facility control subsystem, i.e. the gaming party; u. of_{l_i}And u_{l′_i′}The control quantity of the control devices i and i 'in the control subsystems l and l', namely the strategy adopted by the game party, is represented;the method comprises the steps of representing the optimal control quantity, namely the optimal strategy, of a control device i in a control subsystem l; f. of_lAnd f_l′Representing the control objective function, i.e. the absolute gain, of the control subsystems l and l'; f. of_l(u_{l_i},u_{l′_i′}) And f_l′(u_{l_i},u_{l′_i′}) Denotes a control quantity u of the control devices i and i' in the control subsystems l and l_l,iAnd u_l′,i′Absolute gain of each subsystem;which represents the maximum absolute gain of the entire control system when the control device i in the control subsystem l employs the optimum control quantity.

And S3, solving the obtained optimization function of the target, and obtaining the optimal control quantity of each target function.

S3-1, respectively randomly initializing control sequences u of respective control devices by each control subsystem_{l_i}(0)；

u_{l_i}(0)＝[u_{l_i}(0|0),u_{l_i}(1|0),...,u_{l_i}(M-1|0)],l＝1,2,3

Wherein: m denotes the control time domain, l denotes the control subsystem number, i denotes the number of the control device, u_{l_i}(0|0) denotes the 0 th moment of the control device i in the control sub-system lRandom preset value of control quantity u_{l_i}(M-1|0) represents a random preset value of the control quantity of the control equipment i at the moment 0 in the control subsystem I at the moment M-1;

s3-2, the control subsystems communicate with each other and respectively control the control sequence u of the time k_{l_i,q}(k) Sending the data to other subsystems;

u_{l_i,q}(k)＝[u_{l_i}(k|k),u_{l_i}(k+1|k),...,u_{l_i}(k+M-1|k)]^q

wherein: u. of_{l_i}(k + M-1| k) represents a preset value of a control quantity of a control device i at the moment k to the control quantity of the control device i at the moment k + M-1 in a control subsystem l, and q represents the number of times of the control quantity of the control subsystem l and other subsystems at the moment k;

s3-3, after obtaining the control sequence sent by other control subsystems for the q times at the moment k, the control subsystem l obtains the optimal control sequence by the particle swarm optimization according to the own optimization objective function

$u_{l\_i,q}^{*}\left(k\right)={\[u_{l\_i}^{*}\left(k\right|k),u_{l\_i}^{*}(k+1|k),...,u_{l\_i}^{*}(k+M-1|k)\]}^q$

Wherein,the optimal preset value of the control quantity of the control equipment i at the time k to the time k + M-1 in the control subsystem l is represented;

s3-4, for any one of the gambling parties, if q is satisfied at time k, q is equal to q_maxOrThe transmission of the control sequence is stopped and the optimal control sequence is executed1 st control quantity ofThereafter, the process returns to step S4-2 to calculate the optimal control sequence at time k + 1.

Wherein q is_maxA threshold value representing the maximum number of times the control sequence is sent by the control subsystem at a time, representing the improvement in the control sequence.

In the specific implementation:

the test greenhouse of the invention is exemplified by the following conditions: the main body of the four-span plastic greenhouse is of a light steel structure, the span of a single span is 8.0m, the shoulder height is 3.0m, the top height is 5.0m, and the length of the span is 44.0 m. Taking the facility environment as an example, selecting actual measurement outdoor temperature and humidity and indoor temperature and humidity values of 2015 year 1 and last ten days for 5 continuous days as sample data, and identifying parameters of a facility environment mechanism model by adopting a particle swarm algorithm to obtain model parameters shown in table 1:

TABLE 1 parameter identification results for mechanism models

Note: τ is solar radiation transmittance; c_dA flow coefficient for natural ventilation; c_wThe comprehensive wind pressure coefficient of natural ventilation; h is_cThe heat exchange coefficient of indoor air and air passing through the covering material; c_pThe constant pressure specific heat of air. And then, on the basis of the actual model, simulation control is carried out on temperature and humidity environmental factors of the facility environment by combining measured meteorological data of the external environment. In order to verify the effectiveness and reliability of the control algorithm, different external meteorological data are selected every day, and 8 times of simulation tests are carried out on the control algorithm. The other parameter settings involved in the test were as follows: in the model predictive control, the prediction time domain is 5; the control time domain is 2. The particle swarm algorithm is adopted to calculate the optimization problem, and the related parameter values are as follows: the number of particles is 50, the learning factors c1 and c2 are both 1.4962, the weight w is 0.7298, and the iteration number is 100.

To illustrate the effect of the test, a simulation was selected and analyzed, taking spring 2015, 4-month, 21-day (24-hour) as an example. The environment outside the greenhouse is obtained by measuring the temperature, humidity, solar radiation and wind speed outside the greenhouse every 30min by using an automatic weather station Watchdog2900ET of SPECTRUM corporation in the United states. Initial control of facility environmentThe initial temperature and humidity value is set according to requirements, in order to enable the test to be closer to the real environment, the temperature and humidity value acquired at 0:00 moment of the outdoor environment is used as the initial value of facility environment control, namely the initial indoor temperature is 11 ℃, and the initial indoor absolute humidity is 5.76g/m³I.e. a relative humidity of 62%. The control target is to make the indoor temperature reach 27 ℃, and the indoor absolute humidity reach 18.1g/m³I.e. a relative humidity of 80%, while reducing the energy consumption and losses of the equipment.

In conclusion, the invention researches the multi-target control problem in the facility environment, provides an application framework of the distributed model predictive control in the facility environment aiming at the nonlinearity of the facility environment model, the complexity and the constraint of the objective function, and solves the conflict problems of the three targets of accurate control, minimum energy consumption and minimum equipment opening frequency in the facility environment by adopting a cooperative game method by taking the game theory as reference. Simulation and verification are carried out on an MATLAB platform by taking meteorological data of one day in winter as an example, and comparison is carried out with single-target control and traditional linear weighted multi-target control. Experimental results show that the algorithm provided by the invention has obvious advantages in solving the problem of multi-target conflict, the indoor temperature and humidity can meet the appropriate range required by crop growth, and meanwhile, the energy consumption of control equipment and the equipment loss are greatly reduced.

The parts not involved in the present invention are the same as or can be implemented using the prior art.

Claims (4)

1. A cooperative game-based facility environment multi-objective optimization control method is characterized by comprising the following steps:

s1, three control targets of the facility environment are established: environmental parameters-temperature and humidity control target, energy consumption control target and control equipment physical limit, namely equipment loss control target;

s2, establishing a cooperative game model, and processing three control targets of the facility environment by adopting a multi-target model prediction control method based on the cooperative game model to obtain an optimization function of each control target;

and S3, solving the obtained optimization function of the target, and obtaining the optimal control quantity of each target function.

2. The cooperative game-based facility environment multi-objective optimization control method according to claim 1, wherein the step S1 is specifically as follows:

s1-1, establishing an environment parameter-temperature and humidity control objective function f of the facility₁The description is as follows:

$f_1=min\[\underset{j=1}{\overset{P}{\Σ}}{\left(T_{in}\right(k+j|k)-T_{set})}^2+\underset{j=1}{\overset{P}{\Σ}}{\left(H_{in}\right(k+j|k)-H_{set})}^2\]$

wherein: p represents a prediction time domain, j represents a prediction step length, and k represents the current time; t is_in(k + j | k) represents the room temperature at the future time k + j obtained at the current time k, H_in(k + j | k) represents the indoor humidity at the future time k + j taken at the current time k, T_setAnd H_setRespectively the temperature value and the humidity value which are suitable for the growth of crops;

s1-2, establishing an energy consumption control objective function f for facility control₂The description is as follows:

$f_2=min\[\underset{i=1}{\overset{N}{\Σ}}\underset{j^{\′}=0}{\overset{M-1}{\Σ}}\left(u_i\right(k+j^{\′}|k)/u_{i\_max})/M\]$

wherein N represents controlThe total number of devices, i, represents the number of the current control device; m represents a control time domain, and j' represents a control step size; u. of_i(k + j' | k) represents the optimal control quantity, u, of the future k + j time control device i acquired at the current time k_{i_max}Is the maximum control quantity of the control device i;

s1-3, establishing a device loss control objective function f in the facility₃The description is as follows:

$f_3=min\underset{i=1}{\overset{N}{\Σ}}\left(u_i\right(k+j^{\′}+1|k)-u_i(k+j^{\′}|k\left)\right)/u_{i\_\max }$

wherein: u. of_i(k + j' +1| k) represents the optimum control amount of the control apparatus i at the future time k + j +1 acquired at the current time k.

3. The cooperative game-based facility environment multi-objective optimization control method according to claim 1, wherein the step S2 is specifically as follows:

s2-1, respectively taking three control targets, namely three facility control subsystems, as game parties, and establishing a cooperative game model by adopting the following formula:

$s_l\left(u_l^{*}\right)=\underset{u_l\∈U_l,u_{l^{\′}}\∈U_{l^{\′}}}{max}\[w_{ll}\×{\stackrel{\‾}{s}}_l(u_l,u_{l^{\′}})+\underset{l^{\′}=1(l^{\′}\≠l)}{\overset{m}{\Σ}}{w}_{{\mathrm{ll}}^{\′}}\×{\hat{s}}_{{\mathrm{ll}}^{\′}}(u_l,u_{l^{\′}})\]$

where l and l' denote the number of the gambling parties and m denotes the total number of gambling parties; u shape_lAnd U_l′Set of policies, u, representing gambling parties l and l_lAnd u_l′Indicating the strategy employed by the gambling parties l and l',represents the optimal strategy that the gambling party l can adopt;representing the adoption of policy u by gambling parties l and l', respectively_lAnd u_l′The absolute benefit of the gambling party l,representing the adoption of policy u by gambling parties l and l', respectively_lAnd u_l′The absolute benefit of the betting party l' in time,indicating that betting party l takes an optimal strategyThe maximum absolute income sum which can be obtained by all game parties; w is a_llAnd w_ll′Is a proportionality coefficient, w_llIndicating the extent of non-malfunction of the gambling party l, w_ll′Indicates the degree of cooperation of the gambling party l, and

s2-2, establishing a multi-target model predictive control method based on the cooperative game model, and obtaining an optimized target function of each subsystem in the facility environment as follows:

$S_l\left(u_{l\_i}^{*}\right)=\max \[w_{ll}\×\frac{f_l(u_{l\_i}^{*},u_{l^{\′}\_i^{\′}})}{f_l(u_{l\_i},u_{l^{\′}\_i^{\′}})}+\underset{l^{\′}=1(l^{\′}\≠l)}{\overset{3}{\Σ}}{w}_{{\mathrm{ll}}^{\′}}\×\frac{f_{l^{\′}}(u_{l\_i}^{*},u_{l^{\′}\_i^{\′}})}{f_{l^{\′}}(u_{l\_i},u_{l^{\′}\_i^{\′}})}\]$

wherein: l and l' denote the number of the facility control subsystem, i.e. the gaming party; u. of_{l_i}And u_{l′_i′}The control quantity of the control devices i and i 'in the control subsystems l and l', namely the strategy adopted by the game party, is represented;the method comprises the steps of representing the optimal control quantity, namely the optimal strategy, of a control device i in a control subsystem l; f. of_lAnd f_l′Representing the control objective function, i.e. the absolute gain, of the control subsystems l and l'; f. of_l(u_{l_i},u_{l′_i′}) And f_l′(u_{l_i},u_{l′_i′}) Denotes a control quantity u of the control devices i and i' in the control subsystems l and l_l,iAnd u_l′，i′Absolute gain of each subsystem;which represents the maximum absolute gain of the entire control system when the control device i in the control subsystem l employs the optimum control quantity.

4. The cooperative game-based facility environment multi-objective optimization control method according to claim 1, wherein the step S3 is specifically as follows:

s3-1, respectively randomly initializing control sequences u of respective control devices by each control subsystem_{l_i}(0)；

u_{l_i}(0)＝[u_{l_i}(0|0),u_{l_i}(1|0),...,u_{l_i}(M-1|0)],l＝1,2,3

Wherein: m denotes the control time domain, l denotes the control subsystem number, i denotes the number of the control device, u_{l_i}(0|0) represents a random preset value of the control quantity at the 0 th moment of the control equipment i in the control subsystem l, u_{l_i}(M-1|0) represents a random preset value of the control quantity of the control equipment i at the moment 0 in the control subsystem I at the moment M-1;

s3-2, the control subsystems communicate with each other and respectively control the control sequence u of the time k_{l_i，q}(k) Sending the data to other subsystems;

u_{l_i，q}(k)＝[u_{l_i}(k|k),u_{l_i}(k+1|k),...,u_{l_i}(k+M-1|k)]^q

wherein: u. of_{l_i}(k + M-1| k) represents a preset value of a control quantity of a control device i at the moment k to the control quantity of the control device i at the moment k + M-1 in a control subsystem l, and q represents the number of times of the control quantity of the control subsystem l and other subsystems at the moment k;

s3-3, after obtaining the control sequence sent by other control subsystems for the q times at the moment k, the control subsystem l obtains the optimal control sequence by the particle swarm optimization according to the own optimization objective function

$u_{l\_i,q}^{*}\left(k\right)={\[u_{l\_i}^{*}\left(k\right|k),u_{l\_i}^{*}(k+1|k),...,u_{l\_i}^{*}(k+M-1|k)\]}^q$

Wherein,the optimal preset value of the control quantity of the control equipment i at the time k to the time k + M-1 in the control subsystem l is represented;

s3-4, for any one of the gambling parties, if q is satisfied at time k, q is equal to q_maxOrThe transmission of the control sequence is stopped and the optimal control sequence is executed1 st control quantity ofThereafter, the process returns to step S4-2 to calculate the optimum control sequence at time k +1,

wherein q is_maxIndicating that the control subsystem is at a certain locationThe maximum value of the number of times the control sequence is transmitted at a time represents the threshold for improvement of the control sequence.