CN105335793A - Industrial structure optimization adjustment method for calculating on the basis of pollutant industry discharge capacity - Google Patents

Abstract

The invention discloses an industrial structure optimization adjustment method for calculating on the basis of pollutant industry discharge capacity. The method carries out the calculation of the allowable discharge capacity on a multi-target nonlinear programming method used for the pollutant industry to obtain enough sample data from the point of view of the time of big data, then, a neural network model is established, and finally, the industrial structure is subjected to the optimization adjustment. The invention provides the industrial structure optimization adjustment method for calculating on the basis of the allowable discharge capacity of the pollutant industry. The method causes an industrial structure optimization adjustment process to skip a complex computation process, is simple and easy in implementation on an aspect of operation, and provides a technical support for the industrial structure optimization adjustment.

Description

The industrial structure calculated based on pollutant industry discharge capacity optimizes and revises method

Technical field

The present invention relates to the technical field that Circum-Bohai Sea 13 city major chemical contaminants arranges extra large overall control, be specifically related to a kind of industrial structure calculated based on pollutant industry discharge capacity and optimize and revise method.

Technical background

The Bohai Sea is the unique semiclosed inland sea of China, area 77,284km ², having unique resources advantage and source, ground advantage, is the important back-up system of Bohai Economic Rim.But the marine eco-environment problems such as along with the continuous quickening of coastal cities industrialization, informationization, urbanization process, immediate offshore area water-quality constantly worsens, Disaster And Prevention Measures of Red Tides takes place frequently, marine ecosystems are unbalance, fishery resources decline are also on the rise.From late 1970s, Marine Environment of Bohai Sea quality is degradating trend generally, especially in the middle and later periods nineties.According to Monitoring and assessment result for many years, the pollution range of Bohai Sea immediate offshore area is expansion trend.Within 1992, the Bohai Sea suffers contaminated area about 20%, and by 1998, the contaminated area in the Bohai Sea about 40%.Through improvement in recent years, 2006, the Bohai Sea did not reach the area about 2.0 ten thousand square kilometres of quality standard for clean oceanic water, accounts for 26% of the Bohai Sea total area, and altogether red tide 11 times occur, accumulative occurring area reaches 2980 square kilometres.Pollute situation still severe, pollute marine site and mainly concentrate on the immediate offshore areas such as Liaodong Wan offshore, Bohai Sea Gulf and Laizhou Wan, major pollutants are inorganic nitrogen, reactive phosphate and petroleum-type etc.The land-sourced pollutants such as industrial waste water, sanitary sewage, industry and house refuse, agricultural chemicals, chemical fertilizer exceed standard in a large number, and to enter marine origin be the basic reason that current Bohai environment conditions worsen fails to be curbed to excess emitters.Because Bohai Sea marine eco-environment problem has become the great restraining factors of Bohai Rim's Sustainable Socioeconomic Development, from the later stage nineties 20th century, State Environmental Protection Administration and National Bureau of Oceanography have successively formulated " blue sea, Bohai Sea action plan " and " Bohai Sea comprehensive improvement planning " respectively, carry out emphatically to comprise major chemical contaminants and arrange the Bohai Sea marine eco-environment quality prevention of extra large overall control, control and improvement etc.

Current, under Scientific Development, Resources for construction economizing type, environmental friendliness shape society, particularly promoting economic development to have become each department and implements Eleventh Five-Year Plan socio-economic development and plan the common principle followed with population, resource, environment coordination.Pollutant is arranged extra large overall control research and is not only China's marine eco-environment protection; the Sustainable Exploitation of ocean resources and realize Eleventh Five-Year Plan society of China and sustainable economic development national objective required; also be the Priority setting (scholar Feng Zuo and Wang Hui, 2001) of Eleventh Five-Year Plan Chinese Sea science and environment science basis research simultaneously.For the coastland rapidly of economic development as Circum-Bohai Sea, as early as possible scientifically and rationally implement pollutant arrange extra large overall control just become promote economy, the most important urgent behave of environment sustainable and healthy development.Meanwhile, National Bureau of Oceanography starts in emphasis bay and the enforcement blue sea, river mouth action plan in the beginning of this century, perfect further in the urgent need to highway network arrangement and computing method; And the pollutant of China arranges extra large overall control research still in the research and probe stage, be difficult to meet the actual demand that pollutant arranges extra large overall control.

And in research method and research means, the mathematical model of China's overall control becomes better and approaching perfection day by day, systematic approach is widely used, mainly contains system simulation method, mathematical programming approach, inputoutput analysis method, system dynamics, large system decomposition method, multi-objective programming method, stochastic programming and fuzzy system theory law of planning etc.In recent years, along with computer technology is widely used in environmental planning, various environmental management information system (EMIS) and large-scale system optimum decision support system (DSS) (DSS) and Geographic Information System (GIS) a large amount of in overall control decision system.These modernization meanses and various mathematical model combine, and are the change of the analysis of system synthesis, evaluation and prediction environmental quality, grasp the transportion and transformation of pollutant in environment objectively, determine that the measure of overall control provides strong support.

Therefore, carry out overall control research in the Bohai Sea and there is important theory significance and realistic meaning, and for the enforcement that China Seas and vicinity pollutant arranges extra large overall control, there is important reference significance.This invention is in view of the advantage of multiple objective function model and Nonlinear programming Model, the multiobjective non linear programming function model combining multiple objective function model and Nonlinear programming Model is utilized to obtain abundant sample data, thus set up out RBF neural model, realizing skipping complicated computing function process, proposing a kind of straightforward procedure for optimizing and revising industrial structure.Method of the present invention will arrange extra large total amount Reduced measure for coastland environmental administration and ocean supervision department formulate mandatory pollutant, and provide scientific basis for realizing being managed by target total amount to capacity total amount management transition; Also provide necessary theoretical foundation and technical support for relevant department formulates Regional Economic Planning, industry restructuring, reasonable disposition industrial layout etc. simultaneously.

Summary of the invention

It is based on having the sustainable principle of fair principle, principle of economic benefit, ecnomics and enviroment and the utilization of resources, with Hai Dinglu that the industrial structure calculated based on pollutant industry discharge capacity of the present invention optimizes and revises method, principle is planned as a whole in sea, river, the multi-target non-linear function plan model of administrative area division principle calculates gained data modeling and coming, then the structure that the method model exports also is rationally reliably.Blowdown load distribution is the final purpose of pollutant emission factor to source, and reasonably, the allowable emission of fair each pollution source of determination is also one of key of land-sourced pollutants overall control.At present, determine the permission discharge capacity of each pollution source and industry structure optimization adjustment has multiple method, but respectively have relative merits.Wherein, the industry method of adjustment that pollutant industry allowable emission based on multiobjective non linear programming function model of the present invention calculates is from the view point of large data age, the problem step of industrial structure optimization is made to only have input and output, directly stride across middle complicated computation process widely, reach simple to operation and don't ignore the effect of validity of result.

The industrial structure allowing discharge capacity to calculate based on pollutant industry of the present invention optimizes and revises method, an input and output problem is simply regarded as by industrial structure being optimized and revised problem, be used for setting up RBF neural model according to the rational sample data of outstanding model generation in the past, optimize and revise for later industry.

Method comprises the following steps:

First be the setting of objective function, environmental capacity of water should to be dispensed in administrative region and every profession and trade in region by target function model under the prerequisite meeting economy, environment and social three coordinated development, and calculates the maximum permission discharge capacity of disposal of pollutants unit.Economic target comprises regional GDP (GDP), water environment treatment expense etc., and environmental index comprises discharge quantity and the emission level (ten thousand yuan of output value discharge intensity) of administrative division or industry waste water and pollutant.Social indicator mainly will meet the coordinated development of population and environment; Then be the setting of constraint condition, constraint condition should based on principle of economic benefit, the fairness doctrine, minimum standard of living guarantee principle, and under the prerequisite meeting economy, environment and social three coordinated development, environmental capacity of water is dispensed to every profession and trade in each administrative division or region.Thus, from the setting content of constraint condition, constraint condition comprises economy substantially, environmental protection, social three aspects are carried out.Wherein economic aspect mainly considers the investment cost of each urban area total output value (GDP), environmental protection; the partition capacity in basin, place, each city, sewage and pollutant emission quantity and discharge water equality are mainly considered in environment aspect, and the development level of population is mainly considered in social aspect.Then be try to achieve sample data according to target function model in constraint function model, set up RBF neural model, optimize and revise for later industry.Specific as follows:

The industrial structure allowing discharge capacity to calculate based on pollutant industry optimizes and revises method, it is characterized in that, the method is the pollutant industry Emission amount calculation method based on multiple objective function nonlinear programming, the allowable emission of pollutant industry is calculated by multiple objective function Nonlinear Programming Method, obtain abundant sample data, then RBF neural model is set up, optimizing and revising for industry by these sample datas.The method process optimized and revised industrial structure is skipped complicated computation process, operationally also simple.

The method is that model can coordinate the relation of economy, environment and re-sources preferably, achieves region industrial structure and optimizes and revises and most optimum distribution of resources based on multiple goal regional water pollution thing Industry Total model of optimizing allocation.The method model allows in discharge capacity calculating more outstanding in pollutant industry at present, the data result selecting this model to calculate also be more meet the sustainable principle of fair principle, principle of economic benefit, ecnomics and enviroment and the utilization of resources, with Hai Dinglu, principle, administrative area division principle this five large principle are planned as a whole in sea, river.It is also relatively reasonable that the such sample data of use like this goes to set up the data that network model exports, reliably, and the application that gears to actual circumstances.

RBF network can approach arbitrary nonlinear function, can the regularity being difficult to resolve in disposal system, there is good generalization ability, and RBF network is partial approximation network, namely certain regional area for the input space only has a few to connect weights impact output, therefore study speed of convergence is than very fast.

Described allows the industrial structure of discharge capacity calculating to optimize and revise the advantage that method combines the method model described in claim 2 and 3 based on pollutant industry, for the structure optimization adjustment of industry provides reliable reasonably technical support.

The industrial structure allowing discharge capacity to calculate based on pollutant industry optimizes and revises a method, and its concrete steps are as follows:

(1) target function model is set up:

① $X\≥\underset{j=1}{\overset{k}{\Σ}}{X}_j,(k=1,2,3,......)$

I is pollutant emission unit, x _ithe pollutant being respectively coastal cities allows discharge capacity, 10 ⁴t/a; (2) constraint condition function model is set up

(2.1) economic benefit restricted model

According to the principle that economic benefit increases, the determination of each urban units maximum permissible release will ensure economic sustainable growth simultaneously, and thus in above formula, GDP will meet certain growth, that is:

GDP _t≥GDP ₀·(1+r _g)^t

GDP in formula ₀for the standard year regional GDP in i city, GDP _tbe the regional GDP of t, refer to the regional GDP planning year herein, unit is 10 ⁴unit/a, r _gfor GDP average growth rate per annum, %;

(2.2) pollutants emission intensity restricted model

α _i≤α _i0

Or ${\mathrm{\α}}_i\≤\stackrel{-}{{\mathrm{\α}}_0}$

α _i0, for the mean value of ten thousand yuan of output value discharge intensity and Bohai Rim ten thousand yuan of output value discharge intensity in standard year i city, unit is t/10 ⁴unit.

Suppose that the discharge of economic growth and pollutant exists certain relation:

α _i＝x _i/GDP _i

X in formula _ifor the pollutant in i city allows discharge capacity, unit is 10 ⁴t/a; GDP _ifor the regional GDP in i city, 10 ⁴unit/a; α _ifor pollutant ten thousand yuan of output value discharge intensity in i city, unit is t/10 ⁴unit, (2.3) population restricted model

P _t≤P ₀·(1+r _p)^t

P ₀for the standard year permanent resident population quantity in i city, P _tfor planning the size of population in year, unit is 10 ⁴people/a.R _pfor population mechanical increase in the urban population rate, ‰.Simultaneously in population and sanitary sewage there is quantitative relationship in the discharge of pollutant:

β _i＝W _Hi/P _i

In formula: β _ifor the pollutants emission intensity per capita in i city, Kg/ people; W _hifor the discharge capacity of pollutant in sanitary sewage, 10 ⁴t/a, P _ifor the size of population, 10 ⁴people/a;

(2.4) environmental investment restricted model

R _i0≤R _i≤R′ _i

R in formula _ifor i urban environment control investment accounts for the ratio of regional GDP, R _i0for standard year environmental investment accounts for the ratio of regional GDP, R ' _ifor maximum environmental investment ratio, unit is %.R ' _ithen according to present situation pollutants emission intensity, basin, place partition capacity, and the ratio-dependent of standard year environmental investment;

For each city, there is inversely prroportional relationship in ratio and the pollutants emission intensity of environmental investment, namely the ratio of environmental investment is larger, and the discharge intensity of pollutant is less, and the ratio of environmental investment is less, and the discharge intensity of pollutant is larger:

R _t/R ₀＝α ₀/α _t

(2.5) Condition of Non-Negative Constrains

α _i≥0；β _i≥0；x _i≥0

(3) obtained the discharge capacity of each pollutant industry permission by the model of step (1), (2) for exporting data, master mould parameter is input data, as the sample data of back network model;

(4) RBF (radial basis function) Establishment of Neural Model

(4.1) radial basis function

RBF neural model should meet all sample training data substantially, namely requires that selected radial basis function is through each training data, i.e. F (X _h)=d ^h, h=n, n are selected radial basis function numbers;

Now select n basis function, the corresponding training data of each basis function, each basis function form is:

namely || dist||=||X-c _i||

(4.2) determine that the forecast model based on radial basis function is:

wherein, x=X _h;

(4.3) determination at radial basis function center

(4.3.1) netinit

N training sample is chosen as cluster centre c in the sample data that multiobjective non linear programming function model before random calculates _i(i=1,2 ..., n);

(4.3.2) Nearest Neighbor Method grouping is pressed in the training sample set of input.

According to all sample data x _hwith center c _ibetween Euclidean distance by x _hbe assigned to each cluster set θ of input amendment _h(h=1,2 ..., n);

(4.3.3) cluster centre is readjusted

Calculate each cluster set θ _hthe mean value of middle training sample, namely new cluster centre c _iif new cluster centre no longer changes, then obtained c _ibe the Basis Function Center that RBF neural is final, otherwise return (4.3.2), the center entering next round solves;

(4.4) weights of radial basis function are asked

The forecast model that the pollutant industry that multiobjective non linear programming function model obtains above allows discharge capacity data x to substitute into (4.2) is then obtained:

Equation can write a Chinese character in simplified form into vector form above: Φ W=d

Wherein

W＝(w ₁,w ₂,…,w _n)

Obvious Φ is scale is the symmetric matrix of n, and has nothing to do with the dimension of x, when Φ can the inverse time, can try to achieve weight vector W=Φ ^-1d;

(4.5) radial basis function type is selected

Select Gaussian function as radial basis function type in the present invention, as follows:

And the x of obviously modeling is herein different, so weight vector W can ask, this Gaussian radial basis function is substituted into the required weight vector W in (4.4);

(4.6) variance of radial basis function is asked

Variance can be solved by following formula:

${\mathrm{\σ}}_i=\frac{c_{max}}{\sqrt{2h}},i=1,2,...,h$

Wherein c _maxfor ultimate range between selected center;

(4.7) finally by the forecast model based on radial basis function of central point, weights, variance substitution (4.2), then the industrial structure that can obtain allowing discharge capacity to calculate based on pollutant industry optimizes and revises model:

$y=\underset{i=1}{\overset{n}{\Σ}}{w}_i\exp (-\frac{1}{2{\mathrm{\σ}}_i^2}||x_h-c_i|{|}^2).$

The present invention's beneficial effect compared with prior art:

The invention provides a kind of industrial structure allowing discharge capacity to calculate based on pollutant industry and optimize and revise method, the method process optimized and revised industrial structure is skipped complicated computation process, operationally also simple, for the structure optimization adjustment of industry provides technical support.

Embodiment

For the Bohai Sea, technical scheme of the present invention is further explained below, but protection scope of the present invention is not by any pro forma restriction of embodiment.

The present embodiment is arranged extra large overall control to Circum-Bohai Sea 13 city major chemical contaminants in the process of the present invention and is calculated.The whole Bohai Sea is broadly divided into Liaodong Wan, Laizhou Wan, Bohai Sea Gulf and Central basin four waters, and environmental capacity is AC1, AC2, AC3 respectively; AC4, whole Bohai environment capacity is EC, Dalian, Yingkou; Panjin, Jinzhou, Huludao City, Qinhuangdao; Tangshan, Tianjin, Cangzhou, Binzhou; Dongying, Weifang, Liaodong Wan front 1 to 5 is given in order in city, 13, Yantai; 6 to 7 give Laizhou Wan, and 8 to 10 give Bohai Sea Gulf, and 11 to 13 give Central basin.

Model parameter

(1) basin partition capacity

For Bohai Sea pollutant marine environmental capacity, Jiang Wensheng etc. are three-dimensional convection-diffusion transport model based on HAMSOM Model Establishment, calculate the minimum water physical migration environmental capacity of Bohai Sea COD, under national I class, II class, III class and IV class sea water quality standard condition, be respectively about 36.59,54.88,73.13 and 91.45 ten thousand t/a.And extrapolate the minimum water physical migration environmental capacity of other pollutants.Under national I class, II class, III class and IV class sea water quality standard condition, DIN is about 3.67,5.48,7.31 and 9.15 ten thousand t/a; PO ₄-P is about 0.28,0.55,0.55 and 0.83 ten thousand t/a; Petroleum hydrocarbon is about 0.92,0.92,5.51 and 9.18 ten thousand t/a; Pb (II) is about 0.018,0.092,0.18 and 0.92 ten thousand t/a; Hg (II) is about 0.00092,0.0037,0.0037 and 0.0092 ten thousand t/a; Cd (II) is about 0.018,0.092,0.18 and 0.18 ten thousand t/a.Li Keqiangs etc. establish Bohai Sea pollutant movement multi-box model in multimedium marine environment, and calculate under national I class, II class, III class and IV class sea water quality standard condition according to the standard self-cleaning volumetric method based on movement multi-box model, Bohai Sea DIN benchmark marine environmental capacity is about 74,95,125 and 1,580,000 t/a respectively; PO ₄-P benchmark marine environmental capacity is about 4.8,7.5,8.4,12.6 ten thousand t/a respectively; Petroleum hydrocarbon benchmark marine environmental capacity is about 9.5,9.5,57 and 950,000 t/a respectively; Pb (II) benchmark marine environmental capacity is about 0.48,2.4,4.8 and 240,000 t/a respectively.Jiang Wensheng etc. according to petroleum hydrocarbon in multimedium marine environment movement box model about the analysis result determining petroleum hydrocarbon marine environmental capacity main movement process, with three-dimensional convection-diffusion transport model for basic framework, by linear superposition method, establish Bohai petroleum hydrocarbon main movement process-three-dimensional hydrodynamic force and transport coupling model.Under national I class, II class, III class and IV class sea water quality standard condition, the minimum marine environmental capacity of Bohai petroleum hydrocarbon is respectively 2.8,2.8,16.9 and 28.2 ten thousand t/a.Again according to total quantity control on emission ratio in minimum water quality standard reference mark place's pollutant levels and corresponding Sea Water of The Bohai Gulf, in conjunction with benchmark marine environmental capacity result of calculation, the minimum marine environmental capacity of corresponding pollutant can be estimated.In the result of calculation that the optimization distribution of this permission discharge capacity is based on main migration-transfer process-three-dimensional hydrodynamic force coupling model.

Table 1 Bohai Sea COD and DIN river basins partition capacity (ten thousand t/a)

(2) society, economy, environmental parameter

First, more than Sea And Bohai Sea Coast, but the littoral city merger in the river emptying into the sea that can impact Bohai Sea ecologic environment enters Sea And Bohai Sea Coast 13 city, and the littoral urban society in river emptying into the sea, environment and'economy situation are in Table 3-4.

After the merger of river alongshore city, obtain model desired parameters according to each The Surroundings in Cities, economic statistics in 2005, in table 3, table 4.

The littoral urban society in table 2 river emptying into the sea, ecnomics and enviroment situation

The littoral urban society in table 2 river emptying into the sea, ecnomics and enviroment situation

Table 3 allows discharge capacity Distribution Optimization Model parameter ()

* city merger is carried out, containing city, upstream, basin; * deducts Huanghai Sea part

Table 4 allows discharge capacity Distribution Optimization Model parameter (two)

Model calculation

Using COD and DIN as the object distributed in the present invention, the fgoalattain function of Matlab is utilized to set up multiobjective non linear programming model.First set up the m file of function needed for objective function and nonlinear programming, then determine target, weight and constraint that is linear, nonlinear inequalities.Do not represent with empty matrix containing linear equality and nonlinear inequalities in this model.The software of model running is MATLAB7.0.

The industrial structure allowing discharge capacity to calculate based on pollutant industry optimizes and revises a method, and its concrete steps are as follows:

(1) target function model is set up:

① $X\≥\underset{j=1}{\overset{k}{\Σ}}{X}_j,(k=1,2,3,4)$

② $X_1\≥\underset{j=1}{\overset{l}{\Σ}}{x}_i,(l=1,2,3,4,5)$

③ $X_2\≥\underset{i=6}{\overset{m}{\Σ}}{x}_i,\left(m=6,7\right)$

④ $X_3\≥\underset{i=8}{\overset{n}{\Σ}}{x}_i,\left(n=8,9,10\right)$

⑤ $X_4\≥\underset{i=11}{\overset{o}{\Σ}}{x}_i,\left(o=11,12,13\right)$

I is pollutant emission unit, and the present invention allows discharge capacity to distribute unit using Circum-Bohai Sea 13 coastal cities as pollutant is used for model.X _ithe pollutant being respectively 13 coastal cities allows discharge capacity, 10 ⁴t/a; X ₁, X ₂, X ₃, X ₄be respectively the permission partition capacity of the Liaohe River, the Luanhe River, Haihe River and the Yellow River four large watershed, 10 ⁴t/a.Wherein Liaohe River Basin comprises Dalian, Yingkou, Panjin, Jinzhou and city, 5, Huludao City, and Luan River Basin comprises Qinhuangdao and city, 2, Tangshan, and Haihe basin comprises Tianjin, Cangzhou, city, 3, Binzhou, and Huanghai Sea basin comprises Dongying, Weifang and city, 3, Yantai.

(2) constraint condition function model is set up

(2.1) economic benefit restricted model

According to the principle that economic benefit increases, the determination of each urban units maximum permissible release will ensure economic sustainable growth simultaneously, and thus in above formula, GDP will meet certain growth, that is:

GDP _t≥GDP ₀·(1+r _g)^t

GDP in formula ₀for the standard year regional GDP in i city, GDP _tbe the regional GDP of t, refer to the regional GDP planning year herein, unit is 10 ⁴unit/a.R _gfor GDP average growth rate per annum, %.

(2.2) pollutants emission intensity restricted model

According to principle that is fair and benefit, carry out the constraint of pollutants emission intensity according to the emission level of the basin partition capacity under the present situation discharge capacity of urban pollutant, certain national sea water quality standard and pollutant.If the present situation total emission volumn of urban pollutant exceedes the basin partition capacity under certain national sea water quality standard in basin, then must to pollutant for discharge intensity retrain.If urban pollutant discharge intensity, namely ten thousand yuan of output value discharge intensity are less than the mean value that Bohai Rim's ten thousand yuan of output values discharge intensity, then plan that the pollutants emission intensity in year can not exceed the emission level of its standard year; If ten thousand yuan, city output value discharge intensity is greater than the mean value of Bohai Rim, then plan that the pollutants emission intensity in year can not exceed the average discharge level of Bohai Rim's standard year, to promote the raising of each municipal pollution technical finesse level.That is:

α _i≤α _i0

Or ${\mathrm{\α}}_i\≤\stackrel{-}{{\mathrm{\α}}_0}$

α _i0, for the mean value of ten thousand yuan of output value discharge intensity and Bohai Rim ten thousand yuan of output value discharge intensity in standard year i city, unit is t/10 ⁴unit.

Suppose that the discharge of economic growth and pollutant exists certain relation:

α _i＝x _i/GDP _i

X in formula _ifor the pollutant in i city allows discharge capacity, unit is 10 ⁴t/a; GDP _ifor the regional GDP in i city, 10 ⁴unit/a; α _ifor pollutant ten thousand yuan of output value discharge intensity in i city, unit is t/10 ⁴unit.Under normal circumstances, along with the development of each ecnomics and enviroment, ten thousand yuan of output value discharge intensity α also can change.

(2.3) population restricted model

Pollutant allows discharge capacity to distribute will ensure the needs that the size of population increases, and the growth of the size of population, the target of each department Eleventh Five-Year Plan defined can not be exceeded.

P _t≤P ₀·(1+r _p)^t

P ₀for the standard year permanent resident population quantity in i city, P _tfor planning the size of population in year, unit is 10 ⁴people/a.R _pfor population mechanical increase in the urban population rate, ‰.Simultaneously in population and sanitary sewage there is quantitative relationship in the discharge of pollutant:

β _i＝W _Hi/P _i

In formula: β _ifor the pollutants emission intensity per capita in i city, Kg/ people; W _hifor the discharge capacity of pollutant in sanitary sewage, 10 ⁴t/a, P _ifor the size of population, 10 ⁴people/a.

(2.4) environmental investment restricted model

R _i0≤R _i≤R′ _i

R in formula _ifor i urban environment control investment accounts for the ratio of regional GDP, R _i0for standard year environmental investment accounts for the ratio of regional GDP, R ' _ifor maximum environmental investment ratio, unit is %.R ' _ithen according to present situation pollutants emission intensity, basin, place partition capacity, and the ratio-dependent of standard year environmental investment.

For each city, there is inversely prroportional relationship in ratio and the pollutants emission intensity of environmental investment, namely the ratio of environmental investment is larger, and the discharge intensity of pollutant is less, and the ratio of environmental investment is less, and the discharge intensity of pollutant is larger:

R _t/R ₀＝α ₀/α _t

(2.5) Condition of Non-Negative Constrains

In the optimizing process of model, may occur allowing a part of parameter to be the situation of negative value.This model mathematically can obtain the maximum value allowing discharge capacity, but can not occur, thus in reality

α _i≥0；β _i≥0；x _i≥0

(3) obtained the discharge capacity of each pollutant industry permission by the model of step (1), (2) for exporting data, master mould parameter is input data, as the sample data of back network model.

(4) RBF (radial basis function) Establishment of Neural Model

(4.1) radial basis function

RBF neural model should meet all sample training data substantially, namely requires that selected radial basis function is through each training data, i.e. F (X _h)=d ^h, h=n, n are selected radial basis function numbers.

Now select n basis function, the corresponding training data of each basis function, each basis function form is:

namely || dist||=||X-c _i||

(4.2) determine that the forecast model based on radial basis function is:

wherein, x=X _h.

(4.3) determination at radial basis function center

(4.3.1) netinit.

N training sample is chosen as cluster centre c in the sample data that multiobjective non linear programming function model before random calculates _i(i=1,2 ..., n).

(4.3.2) Nearest Neighbor Method grouping is pressed in the training sample set of input.

According to all sample data x _hwith center c _ibetween Euclidean distance by x _hbe assigned to each cluster set θ of input amendment _h(h=1,2 ..., n).

(4.3.3) cluster centre is readjusted.

Calculate each cluster set θ _hthe mean value of middle training sample, namely new cluster centre c _iif new cluster centre no longer changes, then obtained c _ibe the Basis Function Center that RBF neural is final, otherwise return (4.3.2), the center entering next round solves.

(4.4) weights of radial basis function are asked

The forecast model that the pollutant industry that multiobjective non linear programming function model obtains above allows discharge capacity data x to substitute into (4.2) is then obtained:

Equation can write a Chinese character in simplified form into vector form above: Φ W=d

Wherein

W＝(w ₁,w ₂,…,w _n)

Obvious Φ is scale is the symmetric matrix of n, and has nothing to do with the dimension of x, when Φ can the inverse time, can try to achieve weight vector W=Φ ^-1d.

(4.5) radial basis function type is selected

For some function, as Gaussian function, unusual S type function, intend many quadratic functions, when the x inputted is different, Φ is exactly reversible.Therefore select Gaussian function as radial basis function type in the present invention, as follows:

And the x of obviously modeling is herein different, so weight vector W can ask, this Gaussian radial basis function is substituted into the required weight vector W in (4.4).

(4.6) variance of radial basis function is asked

Variance can be solved by following formula:

${\mathrm{\σ}}_i=\frac{c_{max}}{\sqrt{2h}},i=1,2,...,h$

Wherein c _maxfor ultimate range between selected center.

(4.7) finally by the forecast model based on radial basis function of central point, weights, variance substitution (4.2), then the industrial structure that can obtain allowing discharge capacity to calculate based on pollutant industry optimizes and revises model:

$y=\underset{i=1}{\overset{n}{\Σ}}{w}_i\exp (-\frac{1}{2{\mathrm{\σ}}_i^2}||x_h-c_i|{|}^2).$

Under country's one, two, three, four sea water quality standards, COD and DIN in each city allows discharge capacity appraising model program as follows:

％＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝

% discharge capacity appraising model-multiobjective non linear programming

functionfgoalattain1

clearall；clc

% determines target

options＝optimset('GoalsExactAchieve',5)；

Goal=[AC1, AC2, AC3, AC4, EC]; Basin partition capacity and target marine environment capacity under % various criterion

% determines the weight of target

weight＝abs(goal)；

The given initial value of %, with standard year data

%x (1) ~ x (13) is discharge capacity, and x (14) ~ x (26) is GDP

X0=[W1, W2, W3, W4, W5, W6, W7, W8, W9, W10, W11, W12, W13 ... % standard year pollutant arranges extra large flux

G ₀1, G ₀2, G ₀3, G ₀4, G ₀5, G ₀6, G ₀7, G ₀8, G ₀9, G ₀10, G ₀11, G ₀12, G ₀13]; % standard year GDP

A=[1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Liaohe River Basin partition capacity

0,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Luan River Basin partition capacity

0,0,0,0,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Haihe basin partition capacity

0,0,0,0,0,0,0,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0; % Huanghe valley partition capacity

1,1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0; % Marine Environment of Bohai Sea capacity

0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0,0,0,0,0,0,0,0,0,0; % Dalian GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0,0,0,0,0,0,0,0,0; % Yingkou GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0,0,0,0,0,0,0,0; % Panjin GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0,0,0,0,0,0,0; % Jinzhou GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0,0,0,0,0,0; % Huludao City GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0,0,0,0,0; % Qinhuangdao GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0,0,0,0; % Tangshan GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0,0,0; % Tianjin GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0,0; % Cangzhou GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0,0; % Binzhou GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0,0; % Dongying GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1,0; % Weifang GDP increases

0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 ,-1; % Yantai GDP increases

1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Dalian population increases

0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Yingkou population increases

0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Panjin population increases

0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Jinzhou population increases

0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Huludao City population increases

0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Qinhuangdao population increases

0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Tangshan population increases

0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Tianjin population increases

0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Cangzhou population increases

0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Binzhou population increases

0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Dongying population increases

0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0; % Weifang population increases

0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0]; % Yantai population increases

b＝[AC1,AC2,AC3,AC4,EC,...

-G _t1 ,-G _t2 ,-G _t3 ,-G _t4 ,-G _t5 ,-G _t6 ,-G _t7 ,-G _t8 ,-G _t9 ,-G _t10 ,-G _t11 ,-G _t12 ,-G _t13 ... % plans year GDP

W ' 1, W ' 2, W ' 3, W ' 4, W ' 5, W ' 6, W ' 7, W ' 8, W ' 9, W ' 10, W ' 11, W ' 12, W ' 13]; % plans year population discharge

Aeq＝[]；

beq＝[]；

% nonnegativity restrictions

lb＝[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]；

ub＝[inf,inf,inf,inf,inf,inf,inf,inf,inf,inf,inf,inf,inf,...

inf,inf,inf,inf,inf,inf,inf,inf,inf,inf,inf,inf,inf]；

[x,fval,attainfactor,exitflag]＝fgoalattain(MyObjFun1,x0,goal,weight,A,b,Aeq,beq,lb,...

ub,MyNonLinconstr1,options)

% multiple objective function

functionf＝MyObjFun1(x)

f(1)＝x(1)+x(2)+x(3)+x(4)+x(5)；

f(2)＝x(6)+x(7)；

f(3)＝x(8)+x(9)+x(10)；

f(4)＝x(11)+x(12)+x(13)；

f(5)＝f(1)+f(2)+f(3)+f(4)；

f＝[f(1)；f(2)；f(3)；f(4)；f(5)]；

% non-linear constrain

%+eps prevents 0/0, and * 10000 is unit conversion

function[C,Ceq]＝MyNonLinconstr1(x)

% discharges strength constraint

% discharges upper intensity limit

C (1)=(x (1)+eps)/(x (14)+eps) * 10000-α 1; % Dalian

C (2)=(x (2)+eps)/(x (15)+eps) * 10000-α 2; % Yingkou

C (3)=(x (3)+eps)/(x (16)+eps) * 10000-α 3; % Panjin

C (4)=(x (4)+eps)/(x (17)+eps) * 10000-α 4; % Jinzhou

C (5)=(x (5)+eps)/(x (18)+eps) * 10000-α 5% Huludao City

C (6)=(x (6)+eps)/(x (19)+eps) * 10000-α 6; % Qinhuangdao

C (7)=(x (7)+eps)/(x (20)+eps) * 10000-α 7; % Tangshan

C (8)=(x (8)+eps)/(x (21)+eps) * 10000-α 8; % Tianjin

C (9)=(x (9)+eps)/(x (22)+eps) * 10000-α 9% Cangzhou

C (10)=(x (10)+eps)/(x (23)+eps) * 10000-α 10; % Binzhou

C (11)=(x (11)+eps)/(x (24)+eps) * 10000-α 11% Dongying

C (12)=(x (12)+eps)/(x (25)+eps) * 10000-α 12; % Weifang

C (13)=(x (13)+eps)/(x (26)+eps) * 10000-α 13; % Yantai

% discharges low intensity limit

C (14)=α ' 1-(x (1)+eps)/(x (14)+eps) * 10000; % Dalian

C (15)=α ' 2-(x (2)+eps)/(x (15)+eps) * 10000; % Yingkou

C (16)=α ' 3-(x (3)+eps)/(x (16)+eps) * 10000; % Panjin

C (17)=α ' 4-(x (4)+eps)/(x (17)+eps) * 10000; % Jinzhou

C (18)=α ' 5-(x (5)+eps)/(x (18)+eps) * 10000; % Huludao City

C (19)=α ' 6-(x (6)+eps)/(x (19)+eps) * 10000; % Qinhuangdao

C (20)=α ' 7-(x (7)+eps)/(x (20)+eps) * 10000; % Tangshan

C (21)=α ' 8-(x (8)+eps)/(x (21)+eps) * 10000; % Tianjin

C (22)=α ' 9-(x (9)+eps)/(x (22)+eps) * 10000; % Cangzhou

C (23)=α ' 10-(x (10)+eps)/(x (23)+eps) * 10000; % Binzhou

C (24)=α ' 11-(x (11)+eps)/(x (24)+eps) * 10000; % Dongying

C (25)=α ' 12-(x (12)+eps)/(x (25)+eps) * 10000; % Weifang

C (26)=α ' 13-(x (13)+eps)/(x (26)+eps) * 10000; % Yantai

Ceq＝[]；

％＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝＝

Result of calculation and discussion

Result of calculation

Based on operational model above, under class country of country one, two, three, four sea water quality standard, the permission discharge capacity result of calculation of COD, DIN is as follows:

(1) COD allows discharge capacity to be calculated as follows table:

Table 5COD allows discharge capacity result of calculation

Unit: ten thousand t/a

(2) DIN allows discharge capacity to be calculated as follows table:

Table 6DIN allows discharge capacity result of calculation

Unit: ten thousand t/a

By the parameter of the multiobjective non linear programming function model of table 3 and table 4 as input item, allow discharge capacity result of calculation as output item using COD and DIN of table 5 and table 6, set up RBF radial basis function neural network model according to step (4).To be as the criterion establishment 13 basis functions with Sea And Bohai Sea Coast 13 cities above, calculate 13 centers, 13 variances and 13 weights, then the industrial structure that can obtain allowing discharge capacity to calculate based on pollutant industry optimizes and revises model:

$y=\underset{i=1}{\overset{13}{\Σ}}{w}_i\exp (-\frac{1}{2{\mathrm{\σ}}_i^2}||x_h-c_i|{|}^2)$

The industrial structure allowing discharge capacity to calculate based on pollutant industry by this optimizes and revises model, for later a large amount of parameters input data, only need be inputted this model, then can obtain the Output rusults predicted, then propose control measure and suggestion to optimizing and revising of industrial structure, this makes whole process simply easy.

In sum, a kind of industrial structure allowing discharge capacity to calculate based on pollutant industry of the present invention optimizes and revises method, based on having the sustainable principle of fair principle, principle of economic benefit, ecnomics and enviroment and the utilization of resources, with Hai Dinglu, principle is planned as a whole in sea, river, the multi-target non-linear function plan model of administrative area division principle calculates gained data modeling and coming, therefore the result that the method model exports also is rationally reliably.The industry method of adjustment that pollutant industry allowable emission based on multiobjective non linear programming function model of the present invention calculates is from the view point of large data age, the problem step of industrial structure optimization is made to only have input and output, directly stride across middle complicated computation process widely, reach simple to operation and don't ignore the effect of validity of result.Namely method of the present invention simply regards an input and output problem as by industrial structure being optimized and revised problem, is used for setting up RBF neural model, optimizes and revises for later industry according to the rational sample data of outstanding model generation in the past.

Claims (1)

1. the industrial structure allowing discharge capacity to calculate based on pollutant industry optimizes and revises a method, it is characterized in that its concrete steps are as follows:

(1) target function model is set up:

① $X\≥\underset{j=1}{\overset{k}{\Σ}}{X}_j,(k=1,2,3,......)$

I is pollutant emission unit, x _ithe pollutant being respectively coastal cities allows discharge capacity, 10 ⁴t/a; (2) constraint condition function model is set up

(2.1) economic benefit restricted model

According to the principle that economic benefit increases, the determination of each urban units maximum permissible release will ensure economic sustainable growth simultaneously, and thus in above formula, GDP will meet certain growth, that is:

GDP _t≥GDP ₀·(1+r _g)^t

GDP in formula ₀for the standard year regional GDP in i city, GDP _tbe the regional GDP of t, refer to the regional GDP planning year herein, unit is 10 ⁴unit/a, r _gfor GDP average growth rate per annum, %;

(2.2) pollutants emission intensity restricted model

α _i≤α _i0

Or ${\mathrm{\α}}_i\≤\stackrel{-}{{\mathrm{\α}}_0}$

α _iO, for the mean value of ten thousand yuan of output value discharge intensity and Bohai Rim ten thousand yuan of output value discharge intensity in standard year i city, unit is t/10 ⁴unit;

Suppose that the discharge of economic growth and pollutant exists certain relation:

α _i＝x _i/GDP _i

X in formula _ifor the pollutant in i city allows discharge capacity, unit is 10 ⁴t/a; GDP _ifor the regional GDP in i city, 10 ⁴unit/a; α _ifor pollutant ten thousand yuan of output value discharge intensity in i city, unit is t/10 ⁴unit;

(2.3) population restricted model

P _t≤P ₀·(1+r _p)^t

P ₀for the standard year permanent resident population quantity in i city, P _tfor planning the size of population in year, unit is 10 ⁴people/a; r _pfor population mechanical increase in the urban population rate, ‰; Simultaneously in population and sanitary sewage there is quantitative relationship in the discharge of pollutant:

β _i＝W _Hi/P _i

In formula: β _ifor the pollutants emission intensity per capita in i city, Kg/ people; W _hifor the discharge capacity of pollutant in sanitary sewage, 10 ⁴t/a, P _ifor the size of population, 10 ⁴people/a;

(2.4) environmental investment restricted model

R _i0≤R _i≤R′ _i

R in formula _ifor i urban environment control investment accounts for the ratio of regional GDP, R _i0for standard year environmental investment accounts for the ratio of regional GDP, R ' _ifor maximum environmental investment ratio, unit is %; R ' _ithen according to present situation pollutants emission intensity, basin, place partition capacity, and the ratio-dependent of standard year environmental investment;

For each city, there is inversely prroportional relationship in ratio and the pollutants emission intensity of environmental investment, namely the ratio of environmental investment is larger, and the discharge intensity of pollutant is less, and the ratio of environmental investment is less, and the discharge intensity of pollutant is larger:

R _t/R ₀＝α ₀/α _t；

(2.5) Condition of Non-Negative Constrains

α _i≥0；β _i≥0；x _i≥0；

(3) obtained the discharge capacity of each pollutant industry permission by the model of step (1), (2) for exporting data, master mould parameter is input data, as the sample data of back network model;

(4) radial basis function neural network model is set up, and radial basis function is called for short RBF

(4.1) radial basis function

RBF neural model should meet all sample training data substantially, namely requires that selected radial basis function is through each training data, i.e. F (X _h)=d ^h, h=n, n are selected radial basis function numbers;

Now select n basis function, the corresponding training data of each basis function, each basis function form is:

i.e. ‖ dist ‖=‖ X-c _i‖

(4.2) determine that the forecast model based on radial basis function is:

wherein, x=X _h;

(4.3) determination at radial basis function center

(4.3.1) netinit

N training sample is chosen as cluster centre c in the sample data that multiobjective non linear programming function model before random calculates _i(i=1,2 ..., n);

(4.3.2) Nearest Neighbor Method grouping is pressed in the training sample set of input

According to all sample data x _hwith center c _ibetween Euclidean distance by x _hbe assigned to each cluster set θ of input amendment _h(h=1,2 ..., n);

(4.3.3) cluster centre is readjusted

Calculate each cluster set θ _hthe mean value of middle training sample, namely new cluster centre c _iif new cluster centre no longer changes, then obtained c _ibe the Basis Function Center that RBF neural is final, otherwise return (4.3.2), the center entering next round solves;

(4.4) weights of radial basis function are asked

The forecast model that the pollutant industry that multiobjective non linear programming function model obtains above allows discharge capacity data x to substitute into (4.2) is then obtained:

Equation can write a Chinese character in simplified form into vector form above: Φ W=d

Wherein

W＝(w ₁,w ₂,...,w _n)

Obvious Φ is scale is the symmetric matrix of n, and has nothing to do with the dimension of x, when Φ can the inverse time, can try to achieve weight vector W=Φ ^-1d;

(4.5) radial basis function type is selected

Select Gaussian function as radial basis function type in the present invention, as follows:

And the x of obviously modeling is herein different, so weight vector W can ask, this Gaussian radial basis function is substituted into the required weight vector W in (4.4);

(4.6) variance of radial basis function is asked

Variance can be solved by following formula:

${\mathrm{\σ}}_i=\frac{c_{max}}{\sqrt{2h}},i=1,2,...,h$

Wherein c _maxfor ultimate range between selected center;

(4.7) finally by the forecast model based on radial basis function of central point, weights, variance substitution (4.2), then the industrial structure that can obtain allowing discharge capacity to calculate based on pollutant industry optimizes and revises model:

$y=\underset{i=1}{\overset{n}{\Σ}}{w}_i\exp (-\frac{1}{2{{\mathrm{\σ}}_i}^2}||x_h-c_i|{|}^2).$