CN105427061A - Improved fish swarm algorithm-based tomato seedling stage photosynthesis optimization regulation and control model, establishment method and application - Google Patents

Abstract

The invention relates to an improved fish swarm algorithm-based tomato seedling photosynthesis optimization regulation and control model. According to the model, with temperature T adopted as an independent variable and photon flux density (PFD) corresponding to a light saturation point adopted as a dependent variable, a model formula is put forward. The invention also discloses an establishment method and application of the model. According to the establishment method of the model, multidimensional data are obtained through utilizing a photosynthetic rate double-factor nested test; a temperature and photon flux density coupled photosynthetic rate multivariate nonlinear regression model is constructed; an improved fish swarm algorithm-based photosynthetic rate model optimization method is designed, so that a light saturation point under different temperature conditions is obtained; and the tomato photosynthetic optimization regulation and control model with the light saturation point adopted as a target value is established. As indicated by the result of a model verification test, the light saturation point under different temperature conditions can be dynamically acquired; the determination coefficient of the actual measurement value and calculated value of the light saturation point is 0.967, and the maximum relative error of the actual measurement value and calculated value of the light saturation point is smaller than 2%, and the regulation and control model has high accuracy and has great significance for improving the regulation and control efficiency of light environment of facilities.

Description

Based on tomato seedling phase photosynthesis Optimum Regulation model and the foundation and application of modified fish-swarm algorithm

Technical field

The invention belongs to reading intelligent agriculture technical field, particularly a kind of tomato seedling phase photosynthesis Optimum Regulation model based on modified fish-swarm algorithm and foundation and application.

Background technology

Tomato is one of chief crop of world's greenhouse production, and cultivated area is up to 450.34 ten thousand hm ², and photosynthetic rate quality directly affects the yield and quality of tomato.There are some researches show that photosynthetic rate is subject to the Environmental Factors such as temperature, gas concentration lwevel, photon flux density, tomato particularly receives publicity as its photosynthetic rate of happiness photosensitiveness crop.The people such as Hu Wenhai research shows that the photosynthetic rate of tomato seedling can reduce under low temperature and poor light; People's results of study such as Zhang Fu deposits show that high temperature stress also can reduce the photosynthetic rate of tomato; The people such as Guo Yong verify in 24 ~ 34 DEG C of Suitable ranges, light saturation point can occur reach or after move, cause the change of photosynthetic rate response curve, make it change with the difference of temperature and light photograph.It is affect one of most important factor of photosynthetic rate that above-mentioned research all shows that temperature and light shines.Therefore, how evaluation temperature and illumination are on the impact of tomato photosynthesis, set up photosynthetic Optimum Regulation model, and the photosynthesis rate improving the crops such as tomato has become arable farming field problem demanding prompt solution.

In recent years, photosynthetic rate model, as the theoretical foundation building photosynthetic Optimum Regulation model, at home and abroad obtains and studies widely.The people such as Li Tian determine the Temperature correction model of Tomato in Greenhouse maximum photosynthesis rate; The people such as Lai Linling have studied the Growth trends of the one dimensional infinite wall tomato variety of spring cultivation in hot house and the relation of temperature of shed and illumination; The people such as JingZhang construct temperature, CO ₂concentration and moisture are to the influence function of tomato photosynthetic rate.But above-mentioned research does not all carry out condition of different temperatures optimizing based on photosynthetic rate model, and photosynthetic Optimum Regulation model should be based upon on the basis of optimizing, therefore design the dynamic optimization method based on light saturation point under condition of different temperatures, become the key setting up photosynthetic Optimum Regulation model.

To obtain under condition of different temperatures light saturation point because genetic algorithm ability of searching optimum is strong but local optimal searching ability is poor based on the photosynthetic Optimum Regulation model dynamic of genetic algorithm, cause this model maximum error to reach 6%.In recent years, the artificial fish-swarm algorithm improved is while avoiding basic artificial fish-swarm algorithm travelling speed slow, there is global optimizing ability and by force, be not easily absorbed in Local Extremum, precision generally higher than features such as genetic algorithms, therefore get the attention, and applied research is carried out in multiple fields of the aspect such as navigation, routing optimality, optimizing scheduling of reservoir, power optimization, and obtain good result at sea.More than research is that the tomato luminous environment goal of regulation and control value model set up based on modified fish-swarm algorithm provides theoretical foundation.More than the dynamic optimization studied as light saturation point provides theoretical foundation, but still there is parameter kind, optimizing condition, type function not parity problem in algorithm design.

Summary of the invention

In order to overcome the shortcoming of above-mentioned prior art, the object of the present invention is to provide a kind of tomato seedling phase photosynthesis Optimum Regulation model based on modified fish-swarm algorithm and foundation and application, based on the non-linear tomato photosynthetic rate model of multiple-factor coupling, obtain the maximum photosynthesis rate under different temperatures and light saturation point, thus foundation take maximum photosynthesis rate as the photosynthetic Optimum Regulation model of tomato of target, for the Optimum Regulation of photosynthetic rate is provided fundamental basis.

To achieve these goals, the technical solution used in the present invention is:

Based on a tomato seedling phase photosynthesis Optimum Regulation model for modified fish-swarm algorithm, with temperature T for independent variable, the photon flux density LSP that light saturation point is corresponding is dependent variable, and model formation is:

$LSP=1212e^{-{\left(\frac{T-29.57}{30.65}\right)}^2}+31.11e^{-{\left(\frac{T-23.41}{3.256}\right)}^2},$ LSP refers to the PFD that light saturation point is corresponding.

Present invention also offers the described method for building up based on the tomato seedling phase photosynthesis Optimum Regulation model of modified fish-swarm algorithm, comprise the steps:

First, carry out testing to obtain test figure, process of the test is as follows:

Batch tomato seedling is adopted nutrient matrix hole plate seedling growth in heliogreenhouse and carries out conventional cultivation management, the cycle is one month, and the tomato seedling that after choosing field planting, growing way is basically identical, healthy and strong is tested, and setting gas concentration lwevel is 300 μ ll ^-1; 6 thermogrades are 16 DEG C, 21 DEG C, 25 DEG C, 29 DEG C, 33 DEG C, 37 DEG C; 10 photon flux density gradients are 0 μm of olm ^-2s ^-1, 50 μm of olm ^-2s ^-1, 100 μm of olm ^-2s ^-1, 200 μm of olm ^-2s ^-1, 400 μm of olm ^-2s ^-1, 600 μm of olm ^-2s ^-1, 800 μm of olm ^-2s ^-1, 1000 μm of olm ^-2s ^-1, 1200 μm of olm ^-2s ^-1, 1500 μm of olm ^-2s ^-1, amount to 60 groups of test conditions, often 6 strain seedling duplicate measurementss, 6 Net Photosynthetic Rate values are all chosen in group test, form test sample collection;

Secondly, temperature T, photon flux density PFD and the photosynthetic rate P of multiple-factor coupling is set up according to test figure _nternary non-linear tomato photosynthetic rate model;

P _n＝f(T,PFD)＝420.1-90.11T+0.1696PFD+7.556T ²-0.01896T·PFD-0.001299PFD ²-0.3099T ³+0.0006919T ²·PFD+1.018×10 ^-5T·PFD ²+7.798×10 ^-8PFD ³+0.00622T ⁴-1.589×10 ^-6T ³·PFD-5.389×10 ^-8T ²·PFD ²+2.031×10 ^-9T·PFD ³-6.654×10 ^-11PFD ⁴-4.891×10 ^-5T ⁵-1.598×10 ^-7T ⁴·PFD+7.865×10 ^-9T ³·PFD ²-2.371×10 ^-11T ²·PFD ³-2.079×10 ^-13T·PFD ⁴+1.71×10 ^-14PFD ⁵

The fitting result coefficient of determination is 0.9929.

Described modified fish-swarm algorithm comprises the nested structure of optimizing condition and specified conditions modified shoal of fish optimizing two parts, wherein, the nested structure of optimizing condition refers to employing nested mode, sets up optimizing gradient within the scope of total temperature, thus completes different Optimization goal function structure; The optimizing of the specified conditions modified shoal of fish carries out optimizing according to specific Optimization goal function, finally realizes the light saturation point optimizing of specified temp.

Described specified conditions modified shoal of fish optimizing detailed process is as follows:

With 2 DEG C for step-length sets up optimizing condition data sample set T=(T ¹, T ²... T ^m), wherein T ^m=T _min+ 2*m, T _minrepresent initial temperature value, m is the integer in [0,10] interval; And complete P in ternary non-linear tomato photosynthetic rate model with this sample intensive data _nthe temperature example of=f (T, PFD), objective function P in searching process _n ^m=f (T ^m, PFD), fitness function F=P _n ^m=f (T ^m, PFD), in formula represent specified temp T in sample set ^munder photosynthetic rate;

Based on stochastic generation initial population, utilize fitness function to calculate adaptation value and complete population evaluation, evaluate when its population and do not meet stop condition, then trigger following operation:

First, before carrying out optimizing, the core visual field of modified fish-swarm algorithm and the dynamic conditioning of step-length, its concrete operation method as shown in the formula:

$\left\{\begin{array}{c}visual=visua{l}_{i-1}\×a+visua{l}_{\min }\\ step={\mathrm{step}}_{i-1}\×a+{\mathrm{step}}_{\min }\end{array}\right.$

In formula: visual represents the visual field of this search Artificial Fish; Step represents the step-length of this search Artificial Fish movement; Visual _i-1represent the visual field of previous search Artificial Fish; Step _i-1represent the step-length of previous search Artificial Fish movement; Visual _minrepresent field range minimum change, step _minrepresent step-length minimum change; A represents adjustment factor.Wherein a=exp (-30 × (t/T _max) ^s), t represents current iteration number of times; T _maxrepresent maximum iteration time rate of change; S represents the integer being more than or equal to 1, and its value directly affects the result of variations of a, and the application selects s=1.

Secondly, adopt general fish-swarm algorithm according to the food concentration of each Artificial Fish in population space, crowding and partner's quantity, select to carry out operation of looking for food, bunch and knock into the back, complete the acquisition of the reposition of Artificial Fish, its concrete grammar is as follows:

Foraging behavior: refer to that the state of setting Artificial Fish current is as x _i, in its sensing range, press following formula Stochastic choice state x _j:

x _j＝x _i+(2rand-1)step

In formula: rand represents a random number, if the food concentration Y of this state _j>Y _i, then following formula is used to complete location updating; Otherwise, the more enterprising row iteration of repetitive completes location updating:

$x_{i++}=x_i+rand\×step\frac{x_j-x_i}{\left|\right|x_j-x_i\left|\right|}$

After iteration exceedes number of attempt, then complete location updating according to carrying out random behavior under formula:

x _i++＝x _i+rand×step

To bunch behavior: be the behavior of Artificial Fish towards the movement of partner center, if artificial fish-swarm center food concentration Y _cwith the Y of current foodstuff concentration _i, and the partner quantity n in present viewing field _fbetween, meet Y _c/ n _f> δ Y _i, Artificial Fish performs according to formula following formula behavior of bunching; Otherwise, perform foraging behavior and complete fish school location.Wherein δ is crowding, be used for limit artificial fish-swarm assemble scale.

$x_{i++}=x_i+rand\·step\frac{x_c-x_i}{\left|\right|x_c-x_i\left|\right|}$

To knock into the back behavior: be that Artificial Fish is towards optimal location partner to the behavior of movement.If artificial fish-swarm current optimal location food concentration Y _gbestwith the Y of current foodstuff concentration _imeet Y _gbestn _f> δ Y _itime, Artificial Fish performs by following formula behavior of knocking into the back.Otherwise Artificial Fish performs foraging behavior.

$x_{i++}=x_i+rand\·step\frac{x_{gbest}-x_i}{\left|\right|x_{gbest}-x_i\left|\right|}$

Finally, contrast the corresponding food concentration in position that different behavior obtains Artificial Fish, the Artificial Fish position choosing food concentration high completes location updating, and is optimized process iterates to the new shoal of fish generated, until the optimizing of maximum photosynthesis rate and light saturation point under completing specified temp.Then again extract the optimizing condition at one group of temperature, repeat said process, until complete the photosynthetic rate optimizing at all temperature.

According to optimizing result, obtain corresponding relation between temperature and light saturation point, then, with temperature T for independent variable, with photon flux density PFD corresponding to light saturation point for dependent variable, utilize the method establishment tomato seedling phase of non-linear regression photosynthetic Optimum Regulation model wherein, the coefficient of determination is 0.9999.

Based on this Model Results, can be the regulation and control algorithm adding warm optically-coupled in existing luminous environment regulator control system to provide fundamental basis, its concrete mode is: based on system real time temperature monitoring information, under utilizing this model realization condition of different temperatures, optimum illumination desired value obtains, thus the suitableeest light filling gauge of the system that realizes is calculated.

Compared with prior art, the invention has the beneficial effects as follows:

1) the photosynthetic rate optimization method based on modified fish-swarm algorithm is proposed, by the dynamic conditioning of the visual field and step-length, the active balance difference of fish-swarm algorithm between global convergence and local convergence, the accurately optimizing fast that algorithm global optimizing ability is strong, speed of searching optimization fast, precision can realize light saturation point under different temperatures higher than genetic algorithm.

2) with the goal of regulation and control value model that non-linear regression method is set up, its coefficient of determination is 0.9999, has good fit effect, can realize being that the goal of regulation and control value of target obtains with light saturation point under different temperatures.

3) correlativity and error analysis are carried out to the model calculation value of light saturation point under 21 different temperatures within the scope of 16 DEG C-36 DEG C and measured value.Result shows, the two highly linear is correlated with, and its coefficient of determination is 0.976, maximum relative error is less than ± and 2%.

The photosynthesis Optimum Regulation model that the present invention proposes can be the regulation and control of tomato luminous environment and provides theoretical foundation, and easily extensible is applied to Different Crop, the photosynthetic Optimum Regulation model of different phase is set up, to improve the photosynthetic capacity of chamber crop.

Accompanying drawing explanation

Fig. 1 is modified fish-swarm algorithm process flow diagram of the present invention.

Fig. 2 regulates index variation profiles under the different rate of change of the present invention.

Fig. 3 is the evolutionary process schematic diagram of photosynthetic rate value under different temperatures of the present invention.

Fig. 4 is the dependency diagram in modelling verification of the present invention between PFD measured value and the analogue value.

Embodiment

Embodiments of the present invention are described in detail below in conjunction with drawings and Examples.

The process of establishing of a kind of tomato seedling phase photosynthesis Optimum Regulation model based on modified fish-swarm algorithm of the present invention is as follows:

1 materials and methods

1.1 test method

Experimental test comprises just test and demonstration test link, examination tomato variety is supplied to be " burr 802 ", selecting 5 homogeneous leaf 1 heart tomato seedling of growing way to be colonizated in fills in the plastic tub of Nutrition Soil, Nutrition Soil is the mellow soil adding the peat composed of rotten mosses and composite fertilizer, in experimental period, apply fertilizer, water etc. manages all conveniently carries out, and does not spray any agricultural chemicals and hormone.

The portable photosynthetic instrument of Li-6400XT type that test adopts LI-COR company to produce, can set and regulate and control the Simultaneously test Net Photosynthetic Rate of leaf room subenvironment.Wherein, thermograde is 16,21,25,29,33,37 DEG C, and photon flux density is 0,50,100,200,400,600,800,1000,1200,1500 (umol/m ²s), light temperature coupled nesting test method is adopted to obtain different gradient temperature and illumination to tomato seedling photosynthetic rate, carry out gross error analysis and filtering, secondly its average is calculated respectively to data after filtering gross error, thus photosynthetic rate sample set corresponding under the different temperatures of formation needed for modeling, illumination combination condition.

1.2 method for establishing model

First the data after utilizing process herein set up the non-linear tomato Net Photosynthetic Rate model of light temperature coupling, and in this, as objective function; Secondly adopt modified fish-swarm algorithm to realize light temperature under Nested conditions based on photosynthetic rate model to be coupled optimizing; Finally set up luminous environment goal of regulation and control value model according to nested optimizing result, its flow process as shown in Figure 1.

1.2.1 tomato photosynthetic rate Optimization goal value function is set up

In Fig. 1, multiple nonlinear regression method is adopted to set up the tomato photosynthetic rate model P of light temperature coupling _n=f (T, PFD), wherein P _n-photosynthetic rate value (umol/m ²s), T-temperature (DEG C), PFD-photon flux density (umol/m ²s); And within the scope of 18-34 DEG C, set up optimizing condition data sample set T=(T ¹, T ²..., T ^m), wherein T ^m=18+ (m-1) × 2 (DEG C), m ∈ [1,9]; Complete model P _n=f (T, PFD) is to the instantiation of temperature, thus the Optimization goal value function F under setting up condition of different temperatures ^m=f (T ^m, PFD).

1.2.2 based on the optimizing algorithm of modified fish-swarm algorithm

On the basis setting up tomato Net Photosynthetic Rate model, propose the photosynthetic rate optimizing algorithm based on modified fish-swarm algorithm herein, optimum photosynthetic rate desired value optimizing under completing specified temp.Due to general artificial fish-swarm algorithm performing foraging behavior, behavior of bunching, knock into the back behavior and random behavior time be all subject to the impact of field range and step-length, field range is larger, the global search of Artificial Fish and convergence capabilities are strong, otherwise the local search ability of Artificial Fish is strong; Step-length is larger, then speed of convergence is faster, but occurs oscillatory occurences sometimes, otherwise then speed of convergence is slower, and solving precision is high.Therefore, the dynamic conditioning herein by the visual field and step-length improves general artificial fish-swarm algorithm, to realize taking into account of speed of searching optimization and low optimization accuracy.

Algorithm in earlier stage, adopts the large visual field and large step-length to search for, realizes the whole rough global search solving territory, effectively improves ability of searching optimum and speed of convergence; In mid-term, the visual field and the step-length of employing diminish fast, improve local search ability, avoid occurring that oscillatory occurences affects speed of convergence; In the later stage, the visual field and step-length are down to minimum, near optimum solution territory, carry out fine search, improve local search ability and optimizing result precision.At a certain temperature in optimum photosynthetic rate desired value searching process, objective function is set as the food concentration of Artificial Fish current location.First the initial shoal of fish of stochastic generation, the state vector of the initial shoal of fish individuality of generation is expressed as X=(x ₁, x ₂..., x _i..., x _n), wherein x _ifor wish optimizing variable photon flux density value PFD; And utilize the Optimization goal value function F under the specified temp obtained ^m, as the food concentration Y of optimizing at this temperature, thus utilize Y ^m=F ^mcalculate food concentration and complete evaluation, evaluate when its shoal of fish and do not meet stop condition, then modified shoal of fish optimizing operation.

1) the core visual field of modified fish-swarm algorithm and the dynamic conditioning of step-length, its formula is such as formula shown in (1).

$\{\begin{array}{c}visual={\mathrm{visual}}_{i-1}\×a+{\mathrm{visual}}_{\min }\\ step={\mathrm{step}}_{i-1}\×a+{\mathrm{step}}_{\min }\end{array}---\left(1\right)$

In formula: the visual field of this search Artificial Fish of visual-; The step-length of this search Artificial Fish movement of step-; Visual _i-1the visual field of-previous search Artificial Fish; Step _i-1the step-length of-previous search Artificial Fish movement; A-adjustment factor; Visual _min-field range minimum change, step _min-step-length minimum change.

From formula (1), under the condition that initial value is determined, the visual field and step-length are determined by adjustment factor a, and linear with a.Therefore, a variation tendency should be consistent in the visual field and step-length, adopts exponential type function to build adjustment factor, shown in (2) herein.

a＝exp(-30×(t/T _max) ^s)(2)

In formula: t-current iteration number of times; T _max-maximum iteration time rate of change; The integer that s-is greater than 1, its value directly affects the result of variations of a.Integer in s chooses [1,10], study the change procedure of a under different rate of change condition, its change curve as shown in Figure 2.

As shown in Figure 2, along with rate of change s increases, global search in early stage proportion increases thereupon; Mid-term, proportion was in first increasing the trend reduced afterwards, and overall variation is little; The proportion that later stage adopts the minimum visual field and step-length to carry out fine search obviously reduces.Because crop photosynthesis Rate Models presents unimodal shape function feature, have stronger global convergence ability, its searching process rate of change should select smaller value, reduces the ratio of global search in early stage, increases local fine search proportion, to improve result precision.

Choose s=1 respectively, 2,3,4 carry out verification experimental verification, and when found that s=1, its group of excellent results are best, and avoid s=2,3, the oscillatory occurences that 4 searching processes occur, therefore select s=1 as rate of change herein.

2) adopt general fish-swarm algorithm according to the food concentration of each Artificial Fish in population space, crowding and partner's quantity, select to carry out operation of looking for food, bunch and knock into the back, complete the acquisition of the reposition of Artificial Fish, its concrete grammar is as follows:

Foraging behavior: refer to that the state of setting Artificial Fish current is as x _i, by formula (2) Stochastic choice state x in its sensing range _j:

x _j＝x _i+(2rand-1)step(2)

In formula: rand-random number, if the food concentration Y of this state _j>Y _i, then formula (3) is used to complete location updating; Otherwise, then repetitive (2) is carried out iteration and is completed location updating.After iteration exceedes number of attempt, then carry out random behavior according to formula (4) and complete location updating.

$x_{i++}=x_i+rand\×step\frac{x_j-x_i}{\left|\right|x_j-x_i\left|\right|}---\left(3\right)$

x _i++＝x _i+rand×step(4)

To bunch behavior: be the behavior of Artificial Fish towards the movement of partner center, if artificial fish-swarm center food concentration Y _cwith the Y of current foodstuff concentration _i, and the partner quantity n in present viewing field _fbetween, meet Y _c/ n _f> δ Y _i, Artificial Fish performs according to formula (5) behavior of bunching.Otherwise, perform foraging behavior and complete fish school location.Wherein δ is crowding, be used for limit artificial fish-swarm assemble scale.

$x_{i++}=x_i+rand\·step\frac{x_c-x_i}{\left|\right|x_c-x_i\left|\right|}---\left(5\right)$

To knock into the back behavior: be that Artificial Fish is towards optimal location partner to the behavior of movement.If artificial fish-swarm current optimal location food concentration Y _gbestwith the Y of current foodstuff concentration _imeet Y _gbestn _f> δ Y _itime, Artificial Fish performs by formula (6) behavior of knocking into the back.Otherwise Artificial Fish performs foraging behavior.

$x_{i++}=x_i+rand\·step\frac{x_{gbest}-x_i}{\left|\right|x_{gbest}-x_i\left|\right|}---\left(6\right)$

3) the corresponding food concentration in position that different behavior obtains Artificial Fish is contrasted, the Artificial Fish position choosing food concentration high completes location updating, and process iterates is optimized to the new shoal of fish generated, until the optimizing of maximum photosynthesis rate and light saturation point under completing specified temp.Then again extract the optimizing condition at one group of temperature, repeat said process, until complete the photosynthetic rate optimizing at all temperature.

1.2.3 luminous environment goal of regulation and control value model

Utilize the method for non-linear regression, based on optimizing result, take temperature as independent variable, light saturation point is dependent variable, build tomato luminous environment goal of regulation and control value model, realize the Dynamic Acquisition of arbitrary temp light saturation point.

2 results and discussion

2.1 photosynthetic rate optimizing results are analyzed

On the basis to the modeling of experimental data non-linear regression, utilize modified shoal of fish method and artificial fish-swarm algorithm to carry out optimizing respectively, obtain optimizing result as shown in Figure 3.

As seen from Figure 3, under different temperatures, artificial fish-swarm algorithm and modified fish-swarm algorithm are in the evolution starting stage, its food concentration is less, through looking for food, knock into the back and the process such as cluster, position constantly upgrades, thus individual food concentration in the new shoal of fish is improved constantly, and best photosynthetic rate value increased with evolutionary generation and increased gradually this generation of tomato; And with evolutionary generation increase, when fish-swarm algorithm produce new individuality approach optimum solution time, its individual food concentration keeps constant substantially, thus completes photosynthetic rate optimizing.Analyze two kinds of fish-swarm algorithm optimizing results to find, the optimizing result that under condition of different temperatures, two kinds of algorithms finally obtain is consistent.Along with temperature raises, photosynthetic rate also rises thereupon, after reaching the maximum along with the continuation of temperature raises its continuous decrease, meets the basic law of crop photosynthesis speed.

From Searching efficiency during evolution, modified fish-swarm algorithm evolutionary rate at different temperatures is all obviously better than original fish-swarm algorithm, use genetic algorithm speed of searching optimization similar to document, searching process can be completed within 8 steps, effectively prevent the problem that artificial fish-swarm algorithm travelling speed is slow, show that the optimization method designed is reasonable herein, can be used for such nonlinear multivariable problem optimizing.

2.2 luminous environment goal of regulation and control value model interpretations of result

The light saturation point that under obtaining different temperatures gradient based on modified fish-swarm algorithm, best photosynthetic rate is corresponding, namely luminous environment goal of regulation and control value optimizing result is at each temperature as shown in table 1.

Table 1

As can be found from Table 1, light saturation point can raise with temperature and raise, after reaching uniform temperature, light saturation point is on a declining curve, it is consistent that itself and the maximum photosynthesis rate shown in Fig. 3 (a)-Fig. 3 (i) vary with temperature trend, with Zhang Fu to deposit etc. people researching high-temperature coerce greenhouse tomato Photosynthetic Characteristics is affected time, under gained different temperatures, light saturation point is consistent with the conclusion of maximum photosynthesis rate variation tendency.Can find from plant physiology analysis further: when temperature is 18 DEG C to about 26 DEG C, along with the rising of temperature, the activity, stomatal conductance etc. of blade Rubisco enzyme obviously rise, photosynthetic capacity raises, and therefore its maximum photosynthesis rate, light saturation point rise rapidly with temperature; When 26 DEG C to 32 DEG C, interval owing to substantially reaching tomato optimum growth temperature, may start obviously to weaken due to the impact of temperature on the activity, stomatal conductance etc. of Rubisco enzyme, maximum photosynthesis rate and light saturation point slowly fluctuate with temperature; When temperature is more than 32 DEG C, crop adopts the mode of closing pore to reduce transpiration, thus minimizing moisture loss, cause crop obviously to weaken CO2 receptivity simultaneously, cause and can decline for the CO2 total amount of photosynthesis metabolism, maximum Photosynthetic rates declines, and in reaction, maximum required photon number decline, light saturation point reduce rapidly, to sum up show that the optimizing result of above-mentioned light saturation point meets crop photosynthesis physiological law.

Based on the above results, with temperature T for independent variable, light saturation point is dependent variable, goal of regulation and control value model is set up such as formula shown in (7) with non-linear regression method, its coefficient of determination is 0.9999, and error term quadratic sum is 1.543, and root-mean-square error is 0.712, show that model has good fit effect, can obtain under different temperatures take light saturation point as the goal of regulation and control value LSP of target.

$LSP=1212e^{-{\left(\frac{T-29.57}{30.65}\right)}^2}+31.11e^{-{\left(\frac{T-23.41}{3.256}\right)}^2}---\left(7\right)$

2.3 modelling verification interpretations of result

Adopt different verification mode, by the comparative analysis of light saturation point measured value and models fitting result, verify accuracy and the adaptability of this model.Test chooses the strain of healthy and strong seedling 21 at random in addition as supplying sample originally, with the portable photosynthetic rate instrument of Li-6400XT type, set and regulate and control leaf room subenvironment, wherein utilizing CO2 injection module (6400-01) to control gas concentration lwevel constant is 300 μ L/L, and utilizing temperature control module to obtain thermograde within the scope of 16 DEG C-36 DEG C is corresponding photoresponse curve under the condition of 1 DEG C.Because temperature control module temperature monitoring precision is ± 0.2 DEG C, in the test under photoresponse curve, therefore adopt the mode of duplicate detection, temperature stabilization to be read is treating the light saturation point measured value under testing temperature obtains 21 temperature.And the light saturation point analogue value under utilizing model to calculate corresponding temperature, carry out correlation analysis to above-mentioned test findings, between light saturation point measured value and the analogue value, relation is as shown in Figure 4.The matched curve coefficient of determination is 0.9760, shows that highly linear is correlated with therebetween.

Carry out error analysis to test findings known, light saturation point measured value and analogue value maximum relative error be less than ± 2.00%, shows that the desired value model set up can realize high precision, the Dynamic Acquisition of light saturation point under condition of different temperatures herein.

Claims (7)

1. based on a tomato seedling phase photosynthesis Optimum Regulation model for modified fish-swarm algorithm, it is characterized in that, with temperature T for independent variable, the photon flux density PFD that light saturation point is corresponding is dependent variable, and model formation is:

$LSP=1212e^{-{\left(\frac{T-29.57}{30.65}\right)}^2}+31.11e^{-{\left(\frac{T-23.41}{3.256}\right)}^2},$ LSP refers to the PFD that light saturation point is corresponding.

2. the method for building up of the tomato seedling phase photosynthesis Optimum Regulation model based on modified fish-swarm algorithm according to claim 1, is characterized in that, comprise the steps:

First, carry out testing to obtain test figure, process of the test is as follows:

Batch tomato seedling is adopted nutrient matrix hole plate seedling growth in heliogreenhouse and carries out conventional cultivation management, the cycle is one month, and the tomato seedling that after choosing field planting, growing way is basically identical, healthy and strong is tested, and setting gas concentration lwevel is 300 μ ll ^-1; 6 thermogrades are 16 DEG C, 21 DEG C, 25 DEG C, 29 DEG C, 33 DEG C, 37 DEG C; 10 photon flux density gradients are 0 μm of olm ^-2s ^-1, 50 μm of olm ^-2s ^-1, 100 μm of olm ^-2s ^-1, 200 μm of olm ^-2s ^-1, 400 μm of olm ^-2s ^-1, 600 μm of olm ^-2s ^-1, 800 μm of olm ^-2s ^-1, 1000 μm of olm ^-2s ^-1, 1200 μm of olm ^-2s ^-1, 1500 μm of olm ^-2s ^-1, amount to 60 groups of test conditions, often 6 strain seedling duplicate measurementss, 6 Net Photosynthetic Rate values are all chosen in group test, form test sample collection;

Secondly, temperature T, photon flux density PFD and the photosynthetic rate P of multiple-factor coupling is set up according to test figure _nternary non-linear tomato photosynthetic rate model;

Then, modified fish-swarm algorithm is adopted to realize light temperature coupling optimizing based on photosynthetic rate model;

Finally, photosynthetic Optimum Regulation model is set up according to optimizing result.

3., according to claim 2 based on the method for building up of the tomato seedling phase photosynthesis Optimum Regulation model of modified fish-swarm algorithm, it is characterized in that, described ternary non-linear tomato photosynthetic rate model is shown below:

P _n＝f(T,PFD)＝420.1-90.11T+0.1696PFD+7.556T ²-0.01896T·PFD-0.001299PFD ²-0.3099T ³+0.0006919T ²·PFD+1.018×10 ^-5T·PFD ²+7.798×10 ^-8PFD ³+0.00622T ⁴-1.589×10 ^-6T ³·PFD-5.389×10 ^-8T ²·PFD ²+2.031×10 ^-9T·PFD ³-6.654×10 ^-11PFD ⁴-4.891×10 ^-5T ⁵-1.598×10 ^-7T ⁴·PFD+7.865×10 ^-9T ³·PFD ²-2.371×10 ^-11T ²·PFD ³-2.079×10 ^-13T·PFD ⁴+1.71×10 ^-14PFD ⁵

The fitting result coefficient of determination is 0.9929.

4. according to claim 3 based on the method for building up of the tomato seedling phase photosynthesis Optimum Regulation model of modified fish-swarm algorithm, it is characterized in that, described modified fish-swarm algorithm comprises the nested structure of optimizing condition and specified conditions modified shoal of fish optimizing two parts, wherein, the nested structure of optimizing condition refers to employing nested mode, set up optimizing gradient within the scope of total temperature, thus complete different Optimization goal function structure; The optimizing of the specified conditions modified shoal of fish carries out optimizing according to specific Optimization goal function, finally realizes the light saturation point optimizing of specified temp.

5. according to claim 4 based on the method for building up of the tomato seedling phase photosynthesis Optimum Regulation model of modified fish-swarm algorithm, it is characterized in that, described specified conditions modified shoal of fish optimizing detailed process is as follows:

With 2 DEG C for step-length sets up optimizing condition data sample set T=(T ¹, T ²... T ^m), wherein T ^m=T _min+ 2*m, T _minrepresent initial temperature value, m is the integer in [0,10] interval; And complete P in ternary non-linear tomato photosynthetic rate model with this sample intensive data _nthe temperature example of=f (T, PFD), objective function P in searching process _n ^m=f (T ^m, PFD), fitness function F=P _n ^m=f (T ^m, PFD), P in formula _n ^mrepresent specified temp T in sample set ^munder photosynthetic rate; Utilize fitness function to calculate adaptation value and complete population evaluation, evaluate when its population and do not meet stop condition, then trigger following operation:

First, before carrying out optimizing, the core visual field of modified fish-swarm algorithm and the dynamic conditioning of step-length, its concrete operation method as shown in the formula:

$\left\{\begin{array}{c}visual={\mathrm{visual}}_{i-1}\×a+{\mathrm{visual}}_{\min }\\ step={\mathrm{step}}_{i-1}\×a+{\mathrm{step}}_{\min }\end{array}\right.$

In formula: visual represents the visual field of this search Artificial Fish; Step represents the step-length of this search Artificial Fish movement; Visual _i-1represent the visual field of previous search Artificial Fish; Step _i-1represent the step-length of previous search Artificial Fish movement; Visual _minrepresent field range minimum change, step _minrepresent step-length minimum change; A represents adjustment factor.Wherein a=exp (-30 × (t/T _max) ^s), t represents current iteration number of times; T _maxrepresent maximum iteration time rate of change, s=1,

Secondly, adopt general fish-swarm algorithm according to the food concentration of each Artificial Fish in population space, crowding and partner's quantity, select to carry out operation of looking for food, bunch and knock into the back, complete the acquisition of the reposition of Artificial Fish, its concrete grammar is as follows:

Foraging behavior: refer to that the state of setting Artificial Fish current is as x _i, in its sensing range, press following formula Stochastic choice state x _j:

x _j＝x _i+(2rand-1)step

In formula: rand represents a random number, if the food concentration Y of this state _j>Y _i, then following formula is used to complete location updating; Otherwise, the more enterprising row iteration of repetitive completes location updating:

$x_{i++}=x_i+rand\×step\frac{x_j-x_i}{\left|\right|x_j-x_i\left|\right|}$

After iteration exceedes number of attempt, then complete location updating according to carrying out random behavior under formula:

x _i++＝x _i+rand×step

To bunch behavior: be the behavior of Artificial Fish towards the movement of partner center, if artificial fish-swarm center food concentration Y _cwith the Y of current foodstuff concentration _i, and the partner quantity n in present viewing field _fbetween, meet Y _c/ n _f> δ Y _i, Artificial Fish performs according to formula following formula behavior of bunching; Otherwise, perform foraging behavior complete fish school location, wherein δ is crowding, be used for limit artificial fish-swarm assemble scale;

$x_{i++}=x_i+rand\·step\frac{x_c-x_i}{\left|\right|x_c-x_i\left|\right|}$

To knock into the back behavior: be Artificial Fish towards optimal location partner to the behavior of movement, if artificial fish-swarm current optimal location food concentration Y _gbestwith the Y of current foodstuff concentration _imeet Y _gbest/ n _f> δ Y _itime, Artificial Fish performs by following formula behavior of knocking into the back, otherwise Artificial Fish performs foraging behavior:

$x_{i++}=x_i+rand\·step\frac{x_{gbest}-x_i}{\left|\right|x_{gbest}-x_i\left|\right|}$

Finally, contrast the corresponding food concentration in position that different behavior obtains Artificial Fish, the Artificial Fish position choosing food concentration high completes location updating, and is optimized process iterates to the new shoal of fish generated, until the optimizing of maximum photosynthesis rate and light saturation point under completing specified temp.Then again extract the optimizing condition at one group of temperature, repeat said process, until complete the photosynthetic rate optimizing at all temperature.

6. according to claim 5 based on the method for building up of the tomato seedling phase photosynthesis Optimum Regulation model of modified fish-swarm algorithm, it is characterized in that, according to optimizing result, obtain corresponding relation between temperature and light saturation point, then, with temperature T for independent variable, with photon flux density PFD corresponding to light saturation point for dependent variable, utilize the method establishment tomato seedling phase of non-linear regression photosynthetic Optimum Regulation model, wherein, the coefficient of determination is 0.9999.

7. according to claim 1 based on the application of the tomato seedling phase photosynthesis Optimum Regulation model of modified fish-swarm algorithm, it is characterized in that, based on system real time temperature monitoring information, under utilizing this model realization condition of different temperatures, optimum illumination desired value obtains, thus the suitableeest light filling gauge of the system that realizes is calculated.