CN106055836A - Multi-target optimization method for casting sequence selection, ranking and casting time policy of continuous casting machine - Google Patents

Abstract

The present invention provides a multi-target optimization method for casting sequence selection, ranking and a casting time policy of a continuous casting machine. The method includes the following steps of establishing a multi-target optimization model, wherein the multi-target optimization model takes minimum total punishment, accumulated metal on the production line and non-effective use amount of high-quality molten iron of production batch plan execution conditions of a steel plant as a target function, and the multi-target optimization model is formed by constraint equations of related processing requirements; acquiring a production batch plan of the steel plant, coding based on casting sequence selection and performing population initialization; decoding based on a main constraint satisfaction method, calculating an adaptability value, and acquiring an initial solution set; performing non-dominated ranking and crowding distance ranking; selecting some individuals in the population as parents; performing crossing and variation on the parents; decoding a calculation result and calculating adaptability; determining an elitist solution set, and calculating a crowding distance and ranking; and outputting the elitist solution set, selecting the most satisfied scheme and transmitting the most satisfied scheme to a steelmaking-continuous casting production operating control system. Through adoption of the multi-target optimization method, a furnace casting period of continuous casting production is controlled stably, the algorithm efficiency is better than that of a traditional non-dominated ranking genetic algorithm and a strength pareto evolutionary algorithm.

Description

Continuous casting unit waters heat and selects, sorts and the multiple-objection optimization of casting time decision-making Method

Technical field

The present invention relates to iron and steel production control technology field, be specifically related to a kind of continuous casting unit water heat select, sequence with The Multipurpose Optimal Method of casting time decision-making.

Background technology

Continuous casting is waited in producing to open and is watered secondary heat composition and casting time decision problem, had both determined that conticaster specifically left and has watered Production plan problem, is also the premise formulating reasonable steel-making continuous casting production schedules, and its core missions are pre-from batch plan Suitable heat is selected in scavenger to wait to open the selected heat (including that heat selects and sequence) in watering time as plan phase conticaster, And determine whether to connect between each heat and water and casting time.At present, for this problem, steel mill substantially relies on artificial experience to carry out Decision-making, science and the effectiveness of the result of decision are difficult to ensure that.It is the most with time decision problem that continuous casting opens the heat watered The Optimal Decision-making problem of target multiple constraint, therefore, multiple-objection optimization modeling and the method for solving of studying this problem have important showing Sincere justice and theory value.

In recent years, the multi-objective problem correlational study about steel smelting-continuous casting Production Scheduling field is concentrated mainly on refining The Production Lot Planning of steel-continuous casting is formulated, and production scheduling aspect.

Having studied the formulation problem for Production Lot Planning Yu steel smelting-continuous casting Multiobjective Scheduling plan is independently Carrying out, the former relates generally to the formulation of casting plan or cast design and the combined optimization method of charging plan, the company of being not related to Heat on casting machine selects, sequence and casting time determine problem, only determines stove number in watering time in batch plan；The latter's Research emphasis is the concrete scheduling method of the steel-making continuous casting operation plan under multiple target, usually assumes that conticaster is respectively waited to open and waters time Order and casting time are known conditions, have avoided waiting to open in executable casting plan on conticaster and have watered heat and casting time Determination be affected by the impact of the factor such as metals resources balance on production line, this greatly simplify realistic problem.This grind respectively There is notable difference between method and the reality Production requirement studied carefully, production management of steel mill had both needed to consider production lot meter comprehensively Draw, need again to water and open the multi-objective problem watered and carry out excellent taking into account on the basis of reality produces constraint group to conticaster simultaneously Changing decision-making, that the most substantially formulates conticaster waters subjob Plan Problem, and using this result of decision as formulating steel-making continuous casting The precondition of production plan scheduling.But, owing to the understanding of such problem importance is short of, and model and the difficulty solved Degree is relatively big, causes being determined by artificial experience by dispatcher in production practices, thus brings substantial amounts of uncertainty to production, Have impact on the realization of the productive target of " orderly, stable, efficient ".

Summary of the invention

In order to overcome defect present in above-mentioned prior art, it is an object of the invention to provide a kind of continuous casting unit and water heat Select, sort the Multipurpose Optimal Method with casting time decision-making.

For realizing the above-mentioned purpose of the present invention, the invention provides a kind of continuous casting unit and water heat selection, sort and water with opening The Multipurpose Optimal Method of time decision-making, it comprises the steps:

S1, controller is connected with the MES data storehouse of steel mill, obtains steel mill MES and plans the Production Lot Planning in pre-scavenger, What described batch plan included being assigned on every casting machine water sub-quantity, water time in steel grade classification, section gauge belonging to each heat Lattice and predetermined opening water the moment；

S2, sets up and overstocks amount of metal, high-quality molten iron with always punishment, the production line of steel mill Production Lot Planning implementation status The object function that non-effective utilization is minimum, described object function is:

Min F={f₁,f₂,f₃} (1)

Wherein,

$f_1=\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\underset{k=1}{\overset{K_i}{\mathrm{\Σ}}}({\mathrm{\ξ}}_{ik(k+1)}^1+{\mathrm{\ξ}}_{ik(k+1)}^2)\·z_{ik}\·z_{i(k+1)}+\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{d}_i\·(K_i-\underset{k=1}{\overset{K_i}{\mathrm{\Σ}}}{z}_{ik})+\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\underset{k^d=1}{\overset{K_i^d}{\mathrm{\Σ}}}\left(\right|x_{{\mathrm{ik}}^d}-{\mathrm{\τ}}_{{\mathrm{ik}}^d}^{date}|\·{\mathrm{\ψ}}_i+y_{{\mathrm{ik}}^d}\·e_i)---\left(2\right)$

f₂=Q_O (3)

$f_3=Q_E\·{\mathrm{\π}}^p+Q_I\·{\mathrm{\δ}}^p-\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\underset{k=1}{\overset{K_i}{\mathrm{\Σ}}}{z}_{ik}\·v_{ik}\·q_i\·\frac{1}{1-\mathrm{\η}}---\left(4\right)$

Q_O=Q_E+Q_IV-Q_C-Q_L-Q_S (5)

$Q_C=\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{\mathrm{rq}}_i+\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\underset{k^d=1}{\overset{K_i^d}{\mathrm{\Σ}}}\left(min\right({\mathrm{\τ}}_e,x_{{\mathrm{ik}}^d}+{\mathrm{\τ}}_{{\mathrm{ik}}^d})-x_{{\mathrm{ik}}^d})\·\mathrm{\ρ}\·{\mathrm{wa}}_{{\mathrm{ik}}^d}\·{\mathrm{ws}}_{{\mathrm{ik}}^d}---\left(6\right)$

$Q_L=Q_C\·\frac{\mathrm{\η}}{1-\mathrm{\η}}---\left(7\right)$

$Q_S=\frac{Q_C\·{\mathrm{\τ}}_A}{({\mathrm{\τ}}_e-{\mathrm{\τ}}_s)\·(1-\mathrm{\η})}+Q_{rcon}---\left(8\right)$

(2) formula represents selected heat mutual steel grade difference rejection penalty and difference expense at date of delivery, does not opens the residue watered The rejection penalty of heat, each heat are opened the most on time and are watered rejection penalty sum；

(3) formula represents that production line overstocks amount of metal Q_o；

(4) formula represents high-quality molten iron utilization not yet in effect；

(5) formula represents overstocked amount of metal, is that metals resources based on production line balances and arranges, is entered ferrum by the plan phase respectively Amount Q_E, stock amount of metal Q on initial production line_IV, continuous-casting steel pouring amount Q_C, metal loss amount Q_L, be conducive to producing the stable end of term raw Produce line safety inventory amount of metal Q_SConstitute；

(6) formula represents that the steel amount of watering of each conticaster is respectively by the steel amount of watering of a upper plan phase legacy tasks with wait to open and water time The steel amount of watering of each heat is constituted；

(7) formula represents and waters the metal loss amount that steel amount is corresponding；

(8) formula represents at production line safety inventory amount of metal, and it needs to add on the basis of production line average inventory amount of metal One random fluctuation demand amount of metal Q_rcon；

Wherein, the concrete meaning of symbol is:

1. the symbol, defined and set:

I: casting machine serial number, i ∈ I, I are conticaster set；

K: the heat sequence number of each casting machine of pre-scavenger, k ∈ K_i,K_ijTime j heat collection is watered for pre-scavenger casting machine i Close, K_iFor whole heat set of the conticaster i of pre-scavenger, k successively produced according to the predetermined casting time of each casting machine；

k^d: wait to open and water heat sequence number,Waiting to open and water heat set for conticaster i,It is Casting machine i makes a reservation for minimum opening and waters stove number；

2., known parameters:

q_ik: the Metal Weight of casting machine i heat k；

v_ik: pre-scavenger casting machine i heat k whether fine quality steel；

wa_ik: the specifications section of pre-scavenger casting machine i heat k；

: casting machine i waits to open and waters heat k^dSpecifications section；

ws_ik: the pulling rate of pre-scavenger casting machine i heat k；

: casting machine i waits to open and waters heat k^dPulling rate；

rq_i: the Metal Weight of legacy tasks on casting machine i；

ρ: molten steel density；

η: metal loss factor；

τ_s、τ_e: plan start time phase, finish time；

τ_A: the whole process average stream time；

: casting machine i waits to open and waters heat k^dCasting cycle；

: the predetermined casting time of casting machine i heat k；

e_i: casting machine i heat interruption waters failure costs coefficient；

Q_rcon: random fluctuation demand amount of metal；

δ^p、π^p、δ^p': the fine quality steel ratio of legacy tasks, enter the ferrum high-quality molten iron proportion with opening inventory amount of metal；

ψ_i: open on time and water difference cost coefficient；

: the surcharge that between adjacent heat, steel grade difference causes, if steel grade code identical 0, only belong to the big class of same steel grade a₁, belong to different big class a of steel grade₂；

: predetermined casting time difference surcharge,β₁Poor for date of delivery for adjacent heat Different cost coefficient；

d_i: the heat of casting machine i is not chosen as waiting to open watering heat Damage for Detention cost coefficient；

M: penalty factor, removes illegal solution by arranging penalty factor, and M is sufficiently large positive number, is discontented with to guarantee to solve Punishing by sufficiently large fitness value during foot constraint；

: unify coefficient for dimension；

3., decision variable to be solved:

: casting machine i is chosen as waiting to open watering heat k^dOpen and water the moment；

: binary variable, 1 represents that casting machine i waits to open waters heat k^dBreaking with its tight front heat and water, 0 represents and tight front heat Even water；

z_ik: binary variable, in the 1 pre-scavenger of expression, interior certain heat selected conduct of k of casting machine i is waited to open and is watered heat, and 0 represents certain Heat k is the most selected；

S3, sets up preliminary election tank furnace in Production Lot Planning time with waiting to open and waters the restriction relation of heat relation, and metals resources is flat Weighing apparatus related constraint relation, continuous casting installation for casting pot life restriction relation and waiting opens the order between the heat watered and time-constrain is closed System；

S4, selects to carry out encoding and carrying out initialization of population based on heat sequence number；

S5, decodes and calculates fitness value, it is thus achieved that initial disaggregation, and described fitness function is:

$\begin{array}{cc}\min & F=\{f_{{\mathrm{eval}}_1},f_{{\mathrm{eval}}_2},f_{{\mathrm{eval}}_3}\}\end{array}---\left(9\right)$

Wherein,

$f_{{\mathrm{eval}}_1}=f_1---\left(10\right)$

$f_{{\mathrm{eval}}_3}=f_3+M\·max\{Q_D,Q_E\·{\mathrm{\π}}^p+Q_{IV}\·{\mathrm{\δ}}^p\}---\left(12\right)$

(9) formula represents that whole calculating process is in the hope of fitness functionMinima be target；

(10) formula represents fitness function valueEqual to target function value f₁；

(11) formula represents fitness function valueEqual to target function value f₂Amount of metal relation constraint is overstock plus violating Punishment and casting time surmount the punishment sum of plan phase；

(12) formula represents fitness function valueEqual to target function value f₃With plan with high-quality molten iron beyond can supply The caught hell sum of high-quality molten iron；

S6, the solution concentrating described initial solution carries out non-dominated ranking and sorts with crowding distance；

S7, the part selected in step S6 in population is individual as parent；

S8, the parent selecting step S7 is by parents' Shuangzi multiple-spot detection of casting machine segmentation, and takes a little by casting machine segmentation Random variation, it is ensured that karyological character and the true concordance producing heat sequence number feature；

S9, the result after calculating step S8 decodes and calculates fitness, and described fitness function is fitting in step S5 Response function；

S10, determines elite disaggregation, limits and calculates crowding distance individual amount, calculates crowding distance and sequence；

S11, it may be judged whether reach maximum iteration time, if it is, perform step S12, otherwise, performs step S7；

S12, exports elite disaggregation, selects maximum satisfaction scheme by fuzzy optimizing method and opens as continuous casting and water the heat time Decision method；

S13, by maximum satisfaction scheme transmission to steel smelting-continuous casting production run control system, this system according to described Big satisfaction scheme realizes treating on each conticaster is opened the effective production watering the selection of heat, sequence and casting time decision-making Run and control.

The conticaster of the present invention leaves and waters the Multipurpose Optimal Method of heat and time decision-making and water by driving steel mill's conticaster Heat and the analysis of time decision problem, the heat in considering Production Lot Planning waters mutually closing of heat with waiting to open On the basis of the reality influence factor such as system, metals resources balance, continuous casting installation for casting resource situation, heat time sequencing, the company of establishing The Model for Multi-Objective Optimization watering heat with time decision-making is opened in casting, and with non-dominated sorted genetic algorithm (Non-dominated Sorting Genetic Algorithm, NSGAII) based on devise innovatory algorithm (Improved NSGAII, INSGAII) carry out model solution, improve calculating speed and accuracy.

The additional aspect of the present invention and advantage will part be given in the following description, and part will become from the following description Obtain substantially, or recognized by the practice of the present invention.

Accompanying drawing explanation

Above-mentioned and/or the additional aspect of the present invention and advantage are from combining the accompanying drawings below description to embodiment and will become Substantially with easy to understand, wherein:

Fig. 1 is the algorithm flow chart of the inventive method；

Fig. 2 is the coding schematic diagram selected based on heat sequence number in a kind of preferred implementation of the present invention；

Fig. 3 is that in one preferred implementation of the present invention, elite disaggregation blocks schematic diagram；

Fig. 4 is model decision fitness minima evolutionary process in one preferred implementation of the present invention

Fig. 5 is decision-making Gantt chart in one preferred implementation of the present invention；

Fig. 6 is the Gantt chart of manual decision in example shown in Fig. 5.

Detailed description of the invention

Embodiments of the invention are described below in detail, and the example of described embodiment is shown in the drawings, the most from start to finish Same or similar label represents same or similar element or has the element of same or like function.Below with reference to attached The embodiment that figure describes is exemplary, is only used for explaining the present invention, and is not considered as limiting the invention.

Steel mill is generally of and performs the multiple types of dissimilar task (description etc.), the conticaster of polytypic.Raw Producing batch plan is to be watered secondary form and periodically assign by casting machine with predetermined by the productions command centre ERP of top management To the pre-scavenger of plan (being called for short pre-scavenger) of steel mill MES (Manufacturing Execution System, MES), its information Generally include be assigned on every casting machine water sub-quantity, steel grade classification belonging to each heat in watering time, specifications section and predetermined open Watering the moment etc., belong to thick plan, in the present embodiment, specifications section is width dimensions and the gauge of steel.Steel mill Before formulating Steelmaking-Continuous Casting Production Scheduling plan, it is necessary first to carry out continuous casting and open the decision-making watering heat and time.I.e. need root According to estimating into iron, initial stock's amount of metal of production line, continuous casting in batch plan information, one period (being called for short the plan phase) Which heat conduct the reality factors such as equipment task state, select the batch plan of the concrete decision-making each casting machine in pre-scavenger The waiting to open of this plan phase waters heat, and decision-making respectively wait open water heat heat the most previous with it even water and each heat open water time Carve.This heat that is directed to carries out the scheduling executive mode that conticaster casting task arranges, and is more beneficial for playing conticaster Even water ability, it is achieved the multiobjective management requirement of steel mill.

The inventive method can arrange following precondition: 1. can be according to the batch plan information entered in pre-scavenger, by phase With the sequencing of the predetermined casting time of all heats of target casting machine, carry out heat numbering, distinguish on every conticaster with this Heat essential information to be determined；2., in the plan phase, treating on each casting machine is opened and is watered heat and only have same target in pre-scavenger The heat of casting machine is chosen；3. when pre-scavenger internal memory is not when being chosen as opening, in this plan phase, the heat watered, this casting machine can be given over to After tight, the plan phase continues decision-making.

Optimization aim: mainly carrying out in order and angle that the metals resources such as molten iron can effectively utilize from the beneficially production schedule Degree is designed.With the overstocked amount of metal on the violation punishment of Production Lot Planning, production line, high-quality molten iron utilization not yet in effect Be minimised as optimize multiple target.Wherein, the violation to Production Lot Planning punishes that relating generally to JIT requires lower preliminary election The reselecting and sequencing problem of heat in pond, can use for reference group water with JIT (Tang Lixin, Wang Mengguang, Yang Zihou. steel-making one is even Cast for watering the unknown optimum casting plan model of number of times and algorithm [J]. iron and steel, 1997,32 (7): 19-21) payment method Its implementation effect of quantificational description；Overstocked amount of metal on production line embodies the metals resources equilibrium relation in the plan phase, its amount Change expression formula can be described as: [overstocking amount of metal]=[initial line stock amount of metal]+[adding up into iron]-[add up to water steel Amount]-[accumulative metal loss amount]-[end of term safety on line stock amount of metal], safety on line stock's amount of metal is conducive to production Stable；High-quality molten iron utilization not yet in effect reflects steel mill to be needed to be used for producing fine quality steel by limited high-quality molten iron resource Management requirement.

Constraints: mainly include in pre-scavenger heat with wait to open water the mutual relation of heat, metals resources equilibrium relation, Continuous casting installation for casting pot life, heat time sequencing relation etc..

For ease of describing, it is as follows that Definition Model solves the primary symbols related to:

(1) symbol and set

I: casting machine serial number, i ∈ I, I are conticaster set；

J: pre-scavenger waters sequence number, j_i∈J_i,J_iFor pre-scavenger conticaster i water time set；

K: the heat sequence number of each casting machine of pre-scavenger, k ∈ K_i,K_ijTime j heat collection is watered for pre-scavenger casting machine i Close, K_iFor whole heat set of the conticaster i of pre-scavenger, k successively produced according to the predetermined casting time of each casting machine；

k^d: wait to open and water heat sequence number,Waiting to open and water heat set for conticaster i,It is Casting machine i makes a reservation for minimum opening and waters stove number；

(2) known parameters, is unit or value in its bracket,

q_ik: the Metal Weight (t) of casting machine i heat k；

mark_ik: the steel grade code of pre-scavenger casting machine i heat k；

stye_ik: the big class of steel grade of pre-scavenger casting machine i heat k；

v_ik: pre-scavenger casting machine i heat k whether fine quality steel (1 is yes, and 0 is no)；

wa_ik: the specifications section (m of pre-scavenger casting machine i heat k²)；

: casting machine i waits to open and waters heat k^dSpecifications section (m²)；

ws_ik: pulling rate (the m min of pre-scavenger casting machine i heat k^-1)；

: casting machine i waits to open pulling rate (the m min watering heat kd^-1)；

rq_i: the Metal Weight (t) of legacy tasks on casting machine i；

Rgrade_i: the steel grade code of legacy tasks on casting machine i；

rs_i: casting rate (the m min of legacy tasks on casting machine i^-1)；

ra_i: the specifications section (m of legacy tasks on casting machine i²)；

ρ: molten steel density (t m^-3)；

η: metal loss factor (t t^-1)；

τ_s、τ_e: the plan phase starts, finish time (min)；

τ_A: the whole process average stream time (min)；

τ_ik: for the casting cycle (min) of pre-scavenger casting machine i heat k, τ_ik=q_i/ρ·wa_ik·ws_ik；

: casting machine i waits to open and waters heat k^dCasting cycle (min)；

: minimum time interval (min) between casting machine i heat；

: the predetermined casting time (min) of casting machine i heat k；

: casting machine i maximum pot life (min)；

: casting machine i can use the activity duration (min) between watering time the earliest；

e_i: casting machine i heat interruption waters failure costs coefficient (CNY Time^-1)；

Q_E、Q_IV: the plan phase enters iron and stock's amount of metal (t) on initial production line；

Q_rcon: random fluctuation demand amount of metal (t)；

δ^p、π^p、δ^p': the fine quality steel ratio of legacy tasks, enter the ferrum high-quality molten iron proportion with opening inventory amount of metal；

ψ_i: open on time and water difference cost coefficient (CNY Charge^-1)；

: the surcharge that between adjacent heat, steel grade difference causes, if identical 0 (the CNY Charge of steel grade code^-1), Only belong to big class a of same steel grade₁(CNY·Charge^-1), belong to different big class a of steel grade₂(CNY·Charge^-1)；

: predetermined casting time difference surcharge,β₁Poor for date of delivery for adjacent heat Different cost coefficient (CNY Charge^-1)；

d_i: the heat of casting machine i is not chosen as waiting to open watering heat Damage for Detention cost coefficient (CNY Charge^-1)；

(3) decision variable

: casting machine i is chosen as waiting to open watering heat k^dOpen and water the moment (min)；

: binary variable, 1 represents that casting machine i waits to open waters heat k^dBreaking with its tight front heat and water, 0 represents and tight front heat Even water；

z_ik: binary variable, in the 1 pre-scavenger of expression, interior certain heat selected conduct of k of casting machine i is waited to open and is watered heat, and 0 represents certain Heat k is the most selected.

The Multipurpose Optimal Method watering heat with time decision-making opened by the conticaster that the invention provides, as it is shown in figure 1, include Following steps:

S1, controller is connected with the MES data storehouse of steel mill, obtains steel mill MES and plans the Production Lot Planning in pre-scavenger, What wherein batch plan included being assigned on every casting machine water sub-quantity, water time in steel grade classification, section gauge belonging to each heat Lattice and predetermined opening water the moment, select to carry out encoding and carrying out initialization of population based on heat sequence number；

S2, by the overstocked amount of metal violated on total punishment of Production Lot Planning, production line, the utilization not yet in effect of high-quality molten iron That measures is minimised as decision objective, sets up continuous casting and opens the multi-goal optimizing function watering heat with time decision-making, object function equation As follows:

Min F={f₁,f₂,f₃} (1)

Wherein,

$f_1=\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\underset{k=1}{\overset{K_i}{\mathrm{\Σ}}}({\mathrm{\ξ}}_{ik(k+1)}^1+{\mathrm{\ξ}}_{ik(k+1)}^2)\·z_{ik}\·z_{i(k+1)}+\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{d}_i\·(K_i-\underset{k=1}{\overset{K_i}{\mathrm{\Σ}}}{z}_{ik})+\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\underset{k^d=1}{\overset{K_i^d}{\mathrm{\Σ}}}\left(\right|x_{{\mathrm{ik}}^d}-{\mathrm{\τ}}_{{\mathrm{ik}}^d}^{date}|\·{\mathrm{\ψ}}_i+y_{{\mathrm{ik}}^d}\·e_i)---\left(2\right)$

f₂=Q_O (3)

$f_3=Q_E\·{\mathrm{\π}}^p+Q_I\·{\mathrm{\δ}}^p-\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\underset{k=1}{\overset{K_i}{\mathrm{\Σ}}}{z}_{ik}\·v_{ik}\·q_i\·\frac{1}{1-\mathrm{\η}}---\left(4\right)$

Q_O=Q_E+Q_IV-Q_C-Q_L-Q_S (5)

$Q_C=\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{\mathrm{rq}}_i+\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\underset{k^d=1}{\overset{K_i^d}{\mathrm{\Σ}}}\left(min\right({\mathrm{\τ}}_e,x_{{\mathrm{ik}}^d}+{\mathrm{\τ}}_{{\mathrm{ik}}^d})-x_{{\mathrm{ik}}^d})\·\mathrm{\ρ}\·{\mathrm{wa}}_{{\mathrm{ik}}^d}\·{\mathrm{ws}}_{{\mathrm{ik}}^d}---\left(6\right)$

$Q_L=Q_C\·\frac{\mathrm{\η}}{1-\mathrm{\η}}---\left(7\right)$

$Q_S=\frac{Q_C\·{\mathrm{\τ}}_A}{({\mathrm{\τ}}_e-{\mathrm{\τ}}_s)\·(1-\mathrm{\η})}+Q_{rcon}---\left(8\right)$

(2) formula represents selected heat mutual steel grade difference rejection penalty and difference expense at date of delivery, does not opens the residue watered The rejection penalty of heat, each heat are opened the most on time and are watered rejection penalty sum minimum；

(3) formula represents that production line overstocks amount of metal Q_oMinimum；

(4) formula represents that high-quality molten iron utilization not yet in effect is minimum；

(5) formula represents overstocked amount of metal, is that metals resources based on production line balances and arranges, is entered ferrum by the plan phase respectively Amount Q_E, stock amount of metal Q on initial production line_IV, continuous-casting steel pouring amount Q_C, metal loss amount Q_L, be conducive to producing the stable end of term raw Produce line safety inventory amount of metal Q_SConstitute；

(6) formula represents that the steel amount of watering of each conticaster is respectively by the steel amount of watering of a upper plan phase legacy tasks with wait to open and water time The steel amount of watering of each heat is constituted；

(7) formula represents and waters the metal loss amount that steel amount is corresponding；

(8) formula represents at production line safety inventory amount of metal, and it needs at production line averagely on the basis of stock's amount of metal Add a random fluctuation demand amount of metal Q_rcon, wherein, averagely can be according to " average inventory=average at production line stock's amount of metal Unit interval output × average flow time " calculate.

Wherein, for the sequencing problem of selected heat, will during coding and decoding by selected heat sequence number with open The corresponding relation watering the time embodies indirectly.

S3, sets up preliminary election tank furnace in Production Lot Planning time with waiting to open and waters the restriction relation of heat relation, and metals resources is flat Weighing apparatus related constraint relation, continuous casting installation for casting pot life restriction relation and waiting opens the order between the heat watered and time-constrain is closed System；

Constraints when solving is:

1. in Production Lot Planning, preliminary election tank furnace time waters the constraint of heat relation with waiting to open:

$\underset{k=1}{\overset{K_i}{\mathrm{\Σ}}}{z}_{ik}=K_i^d,\∀i\∈\[1,I\]---\left(9\right)$

$K_i^d\≥K_i^n,\∀i\∈\[1,I\]---\left(11\right)$

(9) the heat quantity that in the formula expression plan phase, in pre-scavenger, each conticaster is selected and each casting machine are waited to open and are watered heat Relation between quantity.

(10) formula is for adapting to resource limit condition and to arrange, and represents waiting to open and watering stove number and criticize less than pre-scavenger of each casting machine Gauge draws the minima of total stove number and this casting machine production capacity demand, wherein,Expression rounds up, and mean () represents that calculating is average Value.

(11) formula is arranged for improving tundish and continuous casting installation for casting utilization rate, watering if representing that casting machine is opened, have to be larger than predetermined This casting machine minimum casting furnace number.

2. metals resources balance related constraint:

Q_O≥0 (12)

Q_D≤Q_E·π^p+Q_IV·δ^p’ (13)

$Q_D=({\mathrm{\δ}}^p\·\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{\mathrm{rq}}_i+\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\underset{k^d=1}{\overset{K_i^d}{\mathrm{\Σ}}}{v}_{{\mathrm{ik}}^d}\·(min\left({\mathrm{\τ}}_e,x_{{\mathrm{ik}}^d}+{\mathrm{\τ}}_{{\mathrm{ik}}^d}\right)-x_{{\mathrm{ik}}^d})\·\mathrm{\ρ}\·{\mathrm{wa}}_{{\mathrm{ik}}^d}\·{\mathrm{ws}}_{{\mathrm{ik}}^d})\·1/(1-\mathrm{\η})---\left(14\right)$

(12) production line of formula overstocks amount of metal non-negative and is present to ensure that the production between the plan phase is stable and arranges.

(13) needed for the fine quality steel that formula represents casting, high-quality iron water amount should not exceed opening inventory and enters ferrum with the plan phase High-quality iron water amount sum.

(14) formula represents the high-quality iron water amount Q required for casting fine quality steel_D, respectively by legacy tasks with treat blow-on time Required two parts high-quality iron water amount is constituted.

3. continuous casting installation for casting pot life retrains, and the casting time of chosen heat is no earlier than when can use the earliest of this casting machine Between.

${\mathrm{\τ}}_i^{earliest}\≤x_{{\mathrm{ik}}^d},\∀i\∈\[1,I\],\∀k^d\∈\[1,K_i^d\]---\left(15\right)$

4. wait to open the order between the heat watered and time-constrain:

$y_{{\mathrm{ik}}^d}\left\{\begin{array}{c}=1,{\mathrm{wa}}_{{\mathrm{ik}}^d}\≠{\mathrm{wa}}_{i(k^d-1)},\∀i\∈\[1,I\],\∀k^d\∈\[1,K_i^d\]\\ =0,{\mathrm{wa}}_{{\mathrm{ik}}^d}={\mathrm{wa}}_{i(k^d-1)},\∀i\∈\[1,I\],\∀k^d\∈\[1,K_i^d\]\end{array}\right.---\left(16\right)$

${\mathrm{\τ}}_s\≤x_{{\mathrm{ik}}^d}\≤{\mathrm{\τ}}_e,\∀i\∈\[1,I\],\∀k^d\∈\[2,K_i^d\]---\left(17\right)$

${\mathrm{\τ}}^{\′}+{\mathrm{\τ}}_i^{gap}\·y_{{\mathrm{ik}}^d}\≤x_{{\mathrm{ik}}^d}\≤{\mathrm{\τ}}^{\′}+({\mathrm{\τ}}_e-{\mathrm{\τ}}^{\′})\·y_{{\mathrm{ik}}^d},\∀i\∈\[1,I\],\∀k^d\∈\[1,K_i^d\]---\left(18\right)$

(16) formula is forced disconnected when representing heat discontinuity surface specification difference.

(17) formula represent chosen respectively wait to open water heat casting time and be in the plan phase.

(18) formula represent when wait open water heat and legacy tasks even water time, casting time point takes legacy tasks finish time, τ '=rq_i/ρ·ra_i·rs_i, when non-company waters, take legacy tasks finish time add heat interval to plan the end of term between certain moment,When respectively connecting each other water until blow-on time, this casting time point takes front heat finish time,When non-company waters, take front heat finish time add heat interval to plan the end of term between some time Carve,

S4, selects to carry out encoding and carrying out initialization of population based on heat sequence number.

S5, decodes and calculates fitness value, it is thus achieved that initial disaggregation.Owing to metals resources constraint (12), (13) formula are subject to simultaneously The impact of all decision variables, it is more difficult to accurately express the quantitative relationship between itself and each decision variable span, for prevent because ofThe factors such as value cause partThe value upper limit may exceed plan phase upper limit τ_e, willWith τ_eRelation also include punishment in In construction of function, fitness function is constructed as follows, and utilizes fitness function to solve:

$\begin{array}{cc}\min & F=\{f_{{\mathrm{eval}}_1},f_{{\mathrm{eval}}_2},f_{{\mathrm{eval}}_3}\}\end{array}---\left(19\right)$

$f_{{\mathrm{eval}}_1}=f_1---\left(20\right)$

$f_{{\mathrm{eval}}_3}=f_3+M\·max\{Q_D,Q_E\·{\mathrm{\π}}^p+Q_{IV}\·{\mathrm{\δ}}^p\}---\left(22\right)$

(19) formula represents that whole calculating process is in the hope of fitness functionMinima be target；

(20) formula represents fitness function valueEqual to target function value f₁；

(21) formula represents fitness function valueEqual to target function value f₂Amount of metal relation constraint is overstock plus violating Punishment and casting time surmount the punishment sum of plan phase；

(22) formula represents fitness function valueEqual to target function value f₃With plan with high-quality molten iron beyond can supply The caught hell sum of high-quality molten iron；

M is penalty factor, illegal solution is removed by arranging penalty factor, and the positive number that M is sufficiently large (needs according to concrete Problem specifically determines, choose herein is 100000), with guarantee solution be unsatisfactory for constraint time punishing by sufficiently large fitness value Penalize that (without concrete regulation, as long as can illegal solution be removed by arranging punishment, the fitness value typically paid for should be extremely Few fitness value an order of magnitude being not affected by punishment that is higher than, otherwise, it is difficult to get rid of illegal solution by punishment)；

: unify coefficient for dimension；

Obtain effective z_ik、And after fitness function, according to z_ikObtain with heat sequence number synopsis and respectively wait out The batch plan information of heat, and then calculate and includeDeng the fitness value that can calculate individuality after coefficient.

S6, the solution concentrating initial solution carries out non-dominated ranking and sorts with crowding distance.

S7, the individuality of a part of scale selected in step S6 in population is as parent, in the side of being preferable to carry out of the present invention In formula, can more preferably select 1/2 scale with the individuality of 2/3,1/2,1/3 or 1/4 scale in selected population as parent Individual as parent, it is ensured that calculating is accurately and quickly.

S8, the parent selecting step S7 is by parents' Shuangzi multiple-spot detection of casting machine segmentation, and takes a little by casting machine segmentation Random variation, it is ensured that karyological character and the true concordance producing heat sequence number feature.

S9, the result after calculating step S8 decodes and calculates fitness, and described fitness function is fitting in step S5 Response function.

S10, determines elite disaggregation, limits and calculates crowding distance individual amount, calculates crowding distance and sequence.

S11, it may be judged whether reach maximum iteration time, if it is, perform step S12, otherwise, performs step S7.

S12, exports elite disaggregation, selects maximum satisfaction scheme by fuzzy optimizing method and opens as continuous casting and water the heat time certainly Plan method.

S13, by maximum satisfaction scheme transmission to steel smelting-continuous casting production run control system, this system according to described Big satisfaction scheme realize treating on each conticaster is opened water heat select, sequence and casting time decision-making effectively produce fortune Row controls.

The method of the present invention solves and encodes, at elite disaggregation strategy with the heat serial number gene selected in pre-scavenger In take adjust traditional calculations order, limit calculate crowding distance individual amount corrective measure, finally utilize pareto Solution carries out the method for fuzzy optimizing to determine final optimization pass solution.

As it is shown in figure 1, algorithm is at initial phase, devises a kind of batch plan with waiting to open and water heat relation constraint foot Under, with heat sequence number select (i.e. z_ik) be gene, solve whether to connect between heat by decoding and water (i.e.) and casting time (i.e.) step-by-step processing method reduce the invalid search space of solution；In the genetic manipulation stage of major cycle, use and divide by casting machine Parents' Shuangzi multiple-spot detection of section, and is taken the way of a random variation to guarantee that karyological character produces with true by casting machine segmentation The concordance of heat sequence number feature；In the non-dominated ranking stage of major cycle, premised on the performance not affecting solution, devise with Limiting and calculating crowding distance individual amount is the new method of core, and the method is by only calculating sequence with to include elite disaggregation in direct Relevant individual restrictive practice, is retaining the calculated load of elite solution link alleviating NSGAII, and non-dominated ranking, Elite disaggregation is filled link and is still used the rule in NSGAII；Last solution formation stages use fuzzy optimizing technology from Pareto solves concentration and selects final decision scheme, is beneficial to dispatcher and directly uses.

In the present embodiment, improving search efficiency, coded method is:

S31, total heat number K of each casting machine i in adding up pre-scavenger_i, each heat is pressed the casting time that batch plan is predeterminedSuccessively sequence, and give heat sequence number to successively each stove after sequence, set up each casting machine heat sequence number reference table, in this enforcement In mode, it is ensured that the order of heat sequence number casting time predetermined with batch plan order keeps concordance, with reforwarding after ensureing That calculates is smoothed out.

S32, by selected heat quantitative range, a length of K of stochastic generation in constraint (9-11) formula_iEach casting machine binary system sequence Row, with heat sequence number reference table one_to_one corresponding.Such as, chromosome (0,1,1,0,1) represents K_i=5,Corresponding reference table, 2nd, 3,5 heat sequence numbers are selected as waiting to open watering heat, the 1st, 4 heats the most selected, can be with the 0-1 ratio in casting machine gene section Meet and change in the range of constraint requirements.

S33, the chromogene section of each casting machine is connected and forms a complete chromosome, and randomly generates setting scale Initial population, coding based on the selection of heat sequence number is as shown in Figure 2.

In the present embodiment, z is selected with heat sequence number_ikChromosome for gene may go out during cross and variation Existing illegal solution, for ease of optimizing, selects (i.e. z obtaining with heat sequence number_ik) be gene coded strings after, except needs are repaired non- Method chromosome z_ik, in addition it is also necessary to first pass through decoding according to constraints and solve other two class decision variables: whether connect between heat and water (i.e.) and casting time is (i.e.), and combine the constraints suitable fitness function of structure.Decoding process is as follows:

S41, processes illegal chromosome: the illegal chromosome after cross and variation can be divided into two classes: selected wait to open water stove Secondary sum makes the heat quantitative range upper limit of calculating more than constraint (10) and is less than the heat quantitative range that constraint (11) formula calculates Lower limit, becomes 0 by have more constraint upper limit number 1 at random for the former, will be less than at random for the latter retraining the 0 of lower limit number Becoming 1, the chromosome not violating constraint is constant；

S42, produces to wait to open according to the heat order from left to right occurring 1 in the chromosome of each casting machine and waters heat sequence, right According to the heat feature in reference table and batch plan, produce in conjunction with constraint (16) formula

Water if Step3 treats that certain heat in blow-on time sequence is disconnected with front heat and have earliest available time, by constraint (15), (17) formula determinesOtherwise, determine according to constraint (17-18) formula

In the present embodiment, fitness function is being obtainedWith z_ik、After, according to z_ikWith Heat sequence number synopsis obtains the batch plan information respectively treating blow-on time, and then calculates and includeAfter coefficient The fitness value of every chromosome can be calculated.

In the present embodiment, determine that elite solution diversity method is:

S51, merges parent and child chromosome, dominance relation between definition individuality, gives each individual tax according to dominance relation Giving sequence number grade and sort, producing the non-dominant individual collections of different sequence number grade, in recording each sequence number grade, non-dominant is individual Number.In the present embodiment, between individuality, dominance relation is determined such that

Assume any two solve S1 and S2 for all targets, S1 is respectively less than S2, then we claim S1 arrange S2, if the solution of S1 Do not arranged by other solutions, then S1 is referred to as non-domination solution；If S1 and S2 is to all targets f1, for f2, f3, f1 (s1) < f1 (s2), f2 (s2)<f2 (s2), f3 (s1)>f3 (s2), say, that difficult point of size between target function value, has mutually domination Relation, claiming S1, S2 now is non-domination solution；The set that all non-domination solution are constituted is pareto disaggregation.

S52, according to non-dominant individual amount in elite disaggregation capacity, each sequence number grade, sequence number grade the most suitable Sequence, determines the maximum sequence number that elite disaggregation can accommodate, and according to the elite disaggregation amount of capacity being manually set, determines that it can hold The maximum sequence number received, calculates crowding distance less than or equal to the individual need of this sequence number value, beyond the individuality in this serial number range not Calculate crowding distance not sort, be directly abandoned；

S53, the number summation solved in calculating elite disaggregation and under current non-dominant grade, and judge that the number of described solution is total Whether more than elite disaggregation scale, if it is, perform step S54, if it is not, perform step S55；

S54, calculates the individual crowding distance of current non-dominant grade descending, elite solution is pressed crowding distance from greatly It is sequentially added into elite disaggregation to little order；

S55, calculates the crowding distance under current non-dominant grade and elite solution is added elite disaggregation；

S56, it may be judged whether reach elite disaggregation scale, if reached, then performs such as step S11 in description, if do not had Reach, then make non-dominant grade add 1, perform step S53.

In the present embodiment, elite solution adds the rule of elite disaggregation and is: setting elite disaggregation size, presses wherein Level sequence number order from small to large adds individuality, until elite disaggregation is filled, in same sequence number grade, when meeting sequence phase When same two are individual, the individuality that preferential interpolation crowding distance is big, the individuality beyond elite disaggregation size is abandoned.

The purpose of elite solution strategy is to ensure that excellent parent heat can enter filial generation smoothly.As it is shown on figure 3, it is traditional The step of the elite disaggregation strategy of NSGAII is:

1. dominance relation between definition individuality, gives each individual imparting sequence number grade according to dominance relation and sorts, producing not Non-dominant individual collections with sequence number grade；

2. according to crowding distance between the individuality of all sequence numbers in this set of adjacent body fitness value calculation；

3. set elite disaggregation size, press grade sequence number order from small to large wherein and add individuality, until elite solution Collection is filled, in same sequence number grade, when meet sequence number identical two individual time, preferential add the individuality that crowding distance is big, Individuality beyond elite disaggregation size is abandoned, and this process is also called elite disaggregation and blocks, and elite disaggregation blocks principle such as Fig. 3 institute Show.

NSGAII, during calculating crowding distance, calculates and is ranked a large amount of beyond elite disaggregation capacity, sequence number etc. Level higher and can the individual crowding distance of abandoned non-dominant, when problem scale is compared with big and time iterations is more, time wave Take more prominent.To this end, the improvement of INSGAII or adjustment are as follows:

Step1, in addition to NSGAII step 1. content, records non-dominant individual amount in each sequence number grade；

Step2, according to non-dominant individual amount, sequence number grade in elite disaggregation capacity, each sequence number grade from small to large Sequentially, the maximum sequence number that elite disaggregation can accommodate is determined；

Step3, the sequence number determined according to Step2 calculates the crowding distance in each grade and sorts, and pressing NSGAII step 3. rule fills elite disaggregation.

INSGAII changes crowding distance and elite disaggregation fill order in NSGAII, only calculates and includes elite solution in The crowding distance (as individual in the range of square brackets in Fig. 3) of the part non-dominant individuality that collection is relevant, does not calculate and can be thrown with sequence Abandon and sequence number is more than elite disaggregation by individual (individuality of band No. # in Fig. 3 set B) for the non-dominant of maximum sequence number contained when being filled Crowding distance, for save the calculating time create condition.

In the present embodiment, genetic manipulation includes intersecting and variation, wherein, intersects and is: owing to each casting machine coding section is straight Connect the feature corresponding to Production Lot Planning, if whole chromosome random selecting point is intersected, may occur that heat sequence number is selected in a large number Select the phenomenon that feature is inconsistent with casting machine actual parameter.To this end, use the parents' Shuangzi multiple-spot detection way pressing casting machine segmentation: 1. First take two different complete chromosomes, 2. pressed each casting machine chromosome lengthIt is divided into i section, 3. dyes at each casting machine Randomly select cross point in body section to intersect.

Variation: by each casting machine chromosome lengthIt is divided into i section, in each casting machine chromosome, randomly selects change point: If the value of change point is 0, become 1, otherwise become 0 from 1.

In the present embodiment, the result solved is the elite solution set represented with fitness value, for ease of dispatcher Directly application, is first converted into object function value set by (19-21) formula, more therefrom determines optimum folding by fuzzy optimizing method Middle solution.Fuzzy optimizing method is as follows:

S71, the proportion ω of each individuality in calculating target function value set_(r,m), ω_(r,m)Represent m-th mesh in individual r Offer of tender numerical value proportion,Represent the minimum, of m-th target function value in object function value set respectively Big value:

${\mathrm{\&omega;}}_{(r,m)}=\left\{\begin{array}{cc}1,& f_{(r,m)}\&le;f_{\left(m\right)}^{\min }\\ \frac{f_{\left(m\right)}^{\max }-f_{(r,m)}}{f_{\left(m\right)}^{\max }-f_{\left(m\right)}^{\min }},& f_{\left(m\right)}^{\min }<f_{{\mathrm{eval}}_{(r,m)}}<f_{\left(m\right)}^{\max }\\ 0,& f_{(r,m)}\&GreaterEqual;f_{\left(m\right)}^{\max }\end{array}\right.---\left(23\right)$

S72, satisfaction ω of all individualities of standardization_r, wherein N is the population scale of elite disaggregation；

${\mathrm{\&omega;}}_r=\frac{\underset{m=1}{\overset{4}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_{(r,m)}}{\underset{r=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{m=1}{\overset{4}{\mathrm{\&Sigma;}}}{\mathrm{\&omega;}}_{(r,m)}}---\left(24\right)$

S73, takes the individuality of standardization Maximum SatisfactionHeat and time decision scheme is watered for finally opening z_ik、

Open by certain steel mill's conticaster domestic based on watering the actual production data of heat and time decision-making, carry out model instance Checking, Model suitability and algorithm performance test, carry out the effectiveness of testing model and algorithm.

Modelling verification example: there are 5 conticasters in this factory, opening in production waters heat with time decision-making by the side of artificial experience Formula completes, and is called for short manual decision.With activity duration day for plan the phase, take production actual achievement data as manual decision's result, and with This basis compared as model decision case verification.Producing input data and be shown in Table 1-3 respectively, as space is limited, table 1 is only given and criticizes Gauge is drawn and is watered time the first stove scheduled start tineAdd the most successively with watering time remaining each stove scheduled start tine interior τ_ik。

For test model adaptability and algorithm performance, first, with table 1-3 data as foundation, by changing parameter I, ∑ K_i、 Q_E、τ_AValue form the problem of 4 kinds of different scales, wherein, parameter I=2 represents the casting of the only the 1st, No. 2 casting machines, and I=3 only represents 1st, 2, No. 3 casting machine casting, by that analogy；Secondly, additionally structure three kinds contrast example, by solving 4 respectively with different examples The results contrast planting different scales problem comes the adaptability of test model and the performance of algorithm.The structure purpose of contrast example and spy Levy and be described as follows:

(1) for the effectiveness of the selected multi-objective Algorithm of checking, design is a kind of based on strength Pareto evolutionary algorithm (Strength Pareto Evolutionary Algorithm, SPEAII), its coding and decoding are with herein.

(2) coded system and the elite solution strategy validity of improvement selected by effect simultaneously, based on NSGAII, with Heat serial number gene encodes, genetic manipulation and the elite same document of solution strategy (DEB K, PRATAP A, AGARWAL S, et al.A fast and elitist multi-objective genetic algorithm:NSGA-II[J].IEEE Transactions on Evolutionary Computation, 2002,6 (2): 182-197), it is designated as NSGAII 1.；

(3) it is the effect of effect elite disaggregation improvement strategy, uses coded system herein, genetic manipulation and elite solution strategy 2. document in putting with (2nd), be designated as NSGAII；

The final Pareto disaggregation of each contrast example selects excellent mode identical with this paper, and performance test all takes the flat of 10 operations Average.

Programming with matlab7.0a for platform, each example is at Intel (R) Core (TM) i3-4010U/1.70GHz/ Independent operating in 4.00GB/WIN7 environment, wherein example parameter is set to: Population Size 40, iterations 100, crossover probability 0.8, mutation probability 0.2, additionally, the outside Population Size of SPEAII is 40.

Table 1 batch plan major parameter

Table 2 casting machine major parameter

Table 3 auxiliary parameter

Fig. 4 is the minimum fitness value evolutionary process of the present invention；Table 4 is that the target function value of two kinds of decision modes compares feelings Condition, the result of decision-making the most of the present invention is pareto disaggregation, and the corresponding target function value of sequence number 14 is the fuzzy choosing of pareto disaggregation Excellent result is the most identical with the target function value calculation of model decision；Fig. 5 is corresponding this of No. 14 scheme in table 4 Bright decision scheme Gantt chart；Fig. 6 is manual decision's Gantt chart.

4 two kinds of decision mode target function values of table compare

From Fig. 4-Fig. 6 and table 4, the method for the present invention is conducive to the stability contorting of heat casting cycle.Identical In the case of metals resources amount, because casting machine pulling rate can be set in the range of technological requirement in advance, arrange opening of each heat of each casting machine to water Moment and casting furnace number, effectively prevent casting cycle wave phenomenon.Manual decision waters heat owing to lacking to opening under many casting machines With the means of time-optimized decision-making, after the most artificially decision casting machine is opened and watered, can only pass through frequently to adjust casting machine pulling rate to maintain casting Machine direct casting, causes at least 15 heats casting cycle big ups and downs occur.The frequent fluctuation of heat casting cycle can seriously be made The about raising of slab quality.

The method of the present invention contributes to the management of steel-making continuous casting implementation plan.The 1st of the inventive method, the 2nd object function Being worth excellent compared with manual decision, and the 3rd target function value is slightly worse, reason is that model decision can optimizing in global scope；And it is artificial Though decision-making can be waited on 1#, 2# casting machine to open the variety steel watering heat casting order, even arbitrarily adding non-plan phase by exchanging (heat NaN in Fig. 6) optimizes the 3rd target, but waters the classification of heat, quantity, sequentially because can not accurately determine respectively to wait to open And casting time, result in the 1st, the 2nd target function value and deteriorate.From Gantt chart feature, model decision contributes to it Upper strata batch plan is connected by user with steel mill production plan, in order to manage at steel-making continuous casting implementation plan.

Open for the conticaster in reality production environment and water heat and time decision problem to relate to influence factor many, manually determine The random big phenomenon of plan, the present invention waters heat relation, metals resources with waiting to open considering batch plan preliminary election tank furnace time Balance, continuous casting installation for casting pot life, wait to open and water between heat the constraints such as time sequencing on the basis of, establish violate with batch plan total Punishment, production line are overstock amount of metal, the minimum object function of the non-effective utilization of high-quality molten iron, production technology and organization of production and are wanted Ask to open for the conticaster retrained and water heat and time decision-making Model for Multi-Objective Optimization.

Devise the improvement non-dominated sorted genetic algorithm INSGAII of this model feature applicable.This algorithm is with heat sequence number It is chosen as gene, is reduced the invalid search space of solution by the step-by-step processing method of decoding two other decision variable of generation, adopt Take and adjust tradition elite disaggregation computation sequence, limit the New Measure of calculating sequence non-dominant crowding distance to alleviate reservation elite Solve the calculated load of link, final elite disaggregation is carried out fuzzy optimizing and generates the model optimization solution being easy to intuitivism apprehension, with It is easy to model user of service's reference.

Experimental results shows: model is conducive to producing continuous casting the stability contorting of each heat casting cycle, and favorably The management of practical implementation plan is produced in steel smelting-continuous casting；Algorithm performance test simultaneously shows: the improvement non-dominated ranking of the present invention Genetic algorithm INSGAII more traditional non-dominated sorted genetic algorithm NSGAII and intensity pareto evolution algorithm SPEAII, is asking Solution continuous casting is opened to water in heat time decision-making multi-objective problem higher efficiency.

In the description of this specification, reference term " embodiment ", " some embodiments ", " example ", " specifically show Example " or the description of " some examples " etc. means to combine this embodiment or example describes specific features, structure, material or spy Point is contained at least one embodiment or the example of the present invention.In this manual, to the schematic representation of above-mentioned term not Necessarily refer to identical embodiment or example.And, the specific features of description, structure, material or feature can be any One or more embodiments or example in combine in an appropriate manner.

Although an embodiment of the present invention has been shown and described, it will be understood by those skilled in the art that: not These embodiments can be carried out multiple change in the case of departing from the principle of the present invention and objective, revise, replace and modification, this The scope of invention is limited by claim and equivalent thereof.

Claims (7)

1. a continuous casting unit waters heat selection, sequence and the Multipurpose Optimal Method of casting time decision-making, it is characterised in that bag Include following steps:

S1, controller is connected with the MES data storehouse of steel mill, obtains steel mill MES and plans the Production Lot Planning in pre-scavenger, described What batch plan included being assigned on every casting machine water sub-quantity, water time in steel grade classification belonging to each heat, specifications section and Predetermined opening waters the moment；

S2, sets up the total punishment with steel mill Production Lot Planning implementation status, production line overstocks amount of metal, high-quality molten iron is non-has The object function that effect utilization is minimum, described object function is:

Min F={f₁,f₂,f₃} (1)

Wherein,

$f_1=\underset{i=1}{\overset{I}{\&Sigma;}}\underset{k=1}{\overset{K_i}{\&Sigma;}}({\mathrm{\&xi;}}_{ik(k+1)}^1+{\mathrm{\&xi;}}_{ik(k+1)}^2)\&CenterDot;z_{ik}\&CenterDot;z_{i(k+1)}+\underset{i=1}{\overset{I}{\&Sigma;}}{d}_i\&CenterDot;(K_i-\underset{k=1}{\overset{K_i}{\&Sigma;}}{z}_{ik})+\underset{i=1}{\overset{I}{\&Sigma;}}\underset{k^d=1}{\overset{K_i^d}{\&Sigma;}}\left(\right|x_{{\mathrm{ik}}^d}-{\mathrm{\&tau;}}_{{\mathrm{ik}}^d}^{date}|\&CenterDot;{\mathrm{\&psi;}}_i+y_{{\mathrm{ik}}^d}\&CenterDot;e_i)---\left(2\right)$

f₂=Q_O (3)

$f_3=Q_E\&CenterDot;{\mathrm{\&pi;}}^p+Q_I\&CenterDot;{\mathrm{\&delta;}}^p-\underset{i=1}{\overset{I}{\&Sigma;}}\underset{k=1}{\overset{K_i}{\&Sigma;}}{z}_{ik}\&CenterDot;v_{ik}\&CenterDot;q_i\&CenterDot;\frac{1}{1-\mathrm{\&eta;}}---\left(4\right)$

Q_O=Q_E+Q_IV-Q_C-Q_L-Q_S (5)

$Q_C=\underset{i=1}{\overset{I}{\&Sigma;}}{\mathrm{rq}}_i+\underset{i=1}{\overset{I}{\&Sigma;}}\underset{k^d=1}{\overset{K_i^d}{\&Sigma;}}\left(\min \right({\mathrm{\&tau;}}_e,x_{{\mathrm{ik}}^d}+{\mathrm{\&tau;}}_{{\mathrm{ik}}^d})-x_{{\mathrm{ik}}^d})\&CenterDot;\mathrm{\&rho;}\&CenterDot;{\mathrm{wa}}_{{\mathrm{ik}}^d}\&CenterDot;{\mathrm{ws}}_{{\mathrm{ik}}^d}---\left(6\right)$

$Q_L=Q_C\&CenterDot;\frac{\mathrm{\&eta;}}{1-\mathrm{\&eta;}}---\left(7\right)$

$Q_S=\frac{Q_C\&CenterDot;{\mathrm{\&tau;}}_A}{({\mathrm{\&tau;}}_e-{\mathrm{\&tau;}}_s)\&CenterDot;(1-\mathrm{\&eta;})}+Q_{rcon}---\left(8\right)$

(2) formula represents selected heat mutual steel grade difference rejection penalty and difference expense at date of delivery, does not opens the residue heat watered Rejection penalty, each heat opens the most on time and waters rejection penalty sum；

(3) formula represents that production line overstocks amount of metal Q_o；

(4) formula represents high-quality molten iron utilization not yet in effect；

(5) formula represents overstocked amount of metal, is that metals resources based on production line balances and arranges, is entered iron Q by the plan phase respectively_E, Stock amount of metal Q on initial production line_IV, continuous-casting steel pouring amount Q_C, metal loss amount Q_L, be conducive to producing stable end of term production line Safety inventory amount of metal Q_SConstitute；

(6) formula represent the steel amount of watering of each conticaster respectively by the steel amount of watering of a upper plan phase legacy tasks and wait to open water time in each stove Secondary steel amount of watering is constituted；

(7) formula represents and waters the metal loss amount that steel amount is corresponding；

(8) formula represents at production line safety inventory amount of metal, and it needs to add one on the basis of production line average inventory amount of metal Random fluctuation demand amount of metal Q_rcon；

Wherein, the concrete meaning of symbol is:

1. the symbol, defined and set:

I: casting machine serial number, i ∈ I, I are conticaster set；

K: the heat sequence number of each casting machine of pre-scavenger, k ∈ K_i,K_ijTime j heat set, K is watered for pre-scavenger casting machine i_i For whole heat set of the conticaster i of pre-scavenger, k successively produced according to the predetermined casting time of each casting machine；

k^d: wait to open and water heat sequence number, Waiting to open and water heat set for conticaster i,It is that casting machine i is pre- Fixed minimum opening waters stove number；

2., known parameters:

q_ik: the Metal Weight of casting machine i heat k；

v_ik: pre-scavenger casting machine i heat k whether fine quality steel；

wa_ik: the specifications section of pre-scavenger casting machine i heat k；

Casting machine i waits to open and waters heat k^dSpecifications section；

ws_ik: the pulling rate of pre-scavenger casting machine i heat k；

Casting machine i waits to open and waters heat k^dPulling rate；

rq_i: the Metal Weight of legacy tasks on casting machine i；

ρ: molten steel density；

η: metal loss factor；

τ_s、τ_e: plan start time phase, finish time；

τ_A: the whole process average stream time；

Casting machine i waits to open and waters heat k^dCasting cycle；

The predetermined casting time of casting machine i heat k；

e_i: casting machine i heat interruption waters failure costs coefficient；

Q_rcon: random fluctuation demand amount of metal；

δ^p、π^p、δ^p': the fine quality steel ratio of legacy tasks, enter the ferrum high-quality molten iron proportion with opening inventory amount of metal；

ψ_i: open on time and water difference cost coefficient；

The surcharge that between adjacent heat, steel grade difference causes, if steel grade code identical 0, only belongs to big class a of same steel grade₁, belong to Different big classes a of steel grade₂；

Predetermined casting time difference surcharge,β₁For taking of adjacent heat difference at date of delivery Use coefficient；

d_i: the heat of casting machine i is not chosen as waiting to open watering heat Damage for Detention cost coefficient；

M: penalty factor, removes illegal solution by arranging penalty factor, and M is sufficiently large positive number, to guarantee that solution is unsatisfactory for about Punishing by sufficiently large fitness value during bundle；

Coefficient is unified for dimension；

3., decision variable to be solved:

Casting machine i is chosen as waiting to open watering heat k^dOpen and water the moment；

Binary variable, 1 represents that casting machine i waits to open waters heat k^dBreaking with its most front heat and water, 0 expression connects with tight front heat waters；

z_ik: binary variable, in the 1 pre-scavenger of expression, interior certain heat selected conduct of k of casting machine i is waited to open and is watered heat, and 0 represents certain heat K is the most selected；

S3, sets up preliminary election tank furnace in Production Lot Planning time with waiting to open and waters the restriction relation of heat relation, and metals resources balances phase Close restriction relation, continuous casting installation for casting pot life restriction relation and wait to open the order between the heat watered and time restriction relation；

S4, selects to carry out encoding and carrying out initialization of population based on heat sequence number；

S5, decodes and calculates fitness value, it is thus achieved that initial disaggregation, and described fitness function is:

$\begin{array}{cc}\min & F=\{f_{{\mathrm{eval}}_1},f_{{\mathrm{eval}}_2},f_{{\mathrm{eval}}_3}\}\end{array}---\left(9\right)$

Wherein,

$f_{{\mathrm{eval}}_1}=f_1---\left(10\right)$

$f_{{\mathrm{eval}}_3}=f_3+M\&CenterDot;max\{Q_D,Q_E\&CenterDot;{\mathrm{\&pi;}}^p+Q_{IV}\&CenterDot;{\mathrm{\&delta;}}^p\}---\left(12\right)$

(9) formula represents that whole calculating process is in the hope of fitness functionMinima be target；

(10) formula represents fitness function valueEqual to target function value f₁；

(11) formula represents fitness function valueEqual to target function value f₂Plus violating the punishment overstocking amount of metal relation constraint With the punishment sum that casting time surmounts the plan phase；

(12) formula represents fitness function valueEqual to target function value f₃With plan high-quality molten iron beyond supplying high-quality The caught hell sum of molten iron；

S6, the solution concentrating described initial solution carries out non-dominated ranking and sorts with crowding distance；

S7, the part selected in step S6 in population is individual as parent；

S8, the parent selecting step S7 is by parents' Shuangzi multiple-spot detection of casting machine segmentation, and takes a little random by casting machine segmentation Variation, it is ensured that karyological character and the true concordance producing heat sequence number feature；

S9, the result after calculating step S8 decodes and calculates fitness, and described fitness function is the fitness in step S5 Function；

S10, determines elite disaggregation, limits and calculates crowding distance individual amount, calculates crowding distance and sequence；

S11, it may be judged whether reach maximum iteration time, if it is, perform step S12, otherwise, performs step S7；

S12, exports elite disaggregation, selects maximum satisfaction scheme by fuzzy optimizing method and opens as continuous casting and water the decision-making of heat time Method；

S13, by maximum satisfaction scheme transmission to steel smelting-continuous casting production run control system, this system is according to described severe fullness sensation in the abdomen Meaning degree scheme realizes treating on each conticaster is opened the effective production run watering the selection of heat, sequence and casting time decision-making Control.

2. continuous casting unit as claimed in claim 1 waters heat selection, sequence and the Multipurpose Optimal Method of casting time decision-making, It is characterized in that, in step S3, in Production Lot Planning, preliminary election tank furnace time waters the restriction relation of heat relation with waiting to open is:

$\underset{k=1}{\overset{K_i}{\&Sigma;}}{z}_{ik}=K_i^d,\&ForAll;i\&Element;\&lsqb;1,I\&rsqb;---\left(13\right)$

$K_i^d\&GreaterEqual;K_i^n,\&ForAll;i\&Element;\&lsqb;1,I\&rsqb;---\left(15\right)$

(13) the heat quantity that in the formula expression plan phase, in pre-scavenger, each conticaster is selected and each casting machine are waited to open the number watering heat Relation between amount,

(14) formula be for adapt to resource limit condition and arrange, represent each casting machine wait to open water stove number less than pre-scavenger batch meter Draw the minima of total stove number and this casting machine production capacity demand, wherein,Expression rounds up, and mean () represents calculating meansigma methods,

(15) formula is arranged for improving tundish and continuous casting installation for casting utilization rate, watering if representing that casting machine is opened, have to be larger than predetermined being somebody's turn to do Casting machine minimum casting furnace number；

Metals resources balance related constraint relation is:

Q_O≥0 (16)

Q_D≤Q_E·π^p+Q_IV·δ^p’ (17)

$Q_D=({\mathrm{\&delta;}}^p\&CenterDot;\underset{i=1}{\overset{I}{\&Sigma;}}{\mathrm{rq}}_i+\underset{i=1}{\overset{I}{\&Sigma;}}\underset{k^d=1}{\overset{K_i^d}{\&Sigma;}}{v}_{{\mathrm{ik}}^d}\&CenterDot;(\min \left({\mathrm{\&tau;}}_e,x_{{\mathrm{ik}}^d}+{\mathrm{\&tau;}}_{{\mathrm{ik}}^d}\right)-x_{{\mathrm{ik}}^d})\&CenterDot;\mathrm{\&rho;}\&CenterDot;{\mathrm{wa}}_{{\mathrm{ik}}^d}\&CenterDot;{\mathrm{ws}}_{{\mathrm{ik}}^d})\&CenterDot;1/(1-\mathrm{\&eta;})---\left(18\right)$

(16) formula represents that production line overstocks amount of metal non-negative；

(17) needed for the fine quality steel that formula represents casting, high-quality iron water amount enters the high-quality of ferrum without departing from opening inventory and plan phase Iron water amount sum；

(18) formula represents the high-quality iron water amount Q required for casting fine quality steel_D, respectively by legacy tasks with treat that blow-on time is required Two parts high-quality iron water amount is constituted；

Continuous casting installation for casting pot life restriction relation is:

${\mathrm{\&tau;}}_i^{earliest}\&le;x_{{\mathrm{ik}}^d},\&ForAll;i\&Element;\&lsqb;1,I\&rsqb;,\&ForAll;k^d\&Element;\&lsqb;1,K_i^d\&rsqb;---\left(19\right)$

The casting time being i.e. selected heat is no earlier than the earliest available time of this casting machine；

Wait to open the order between the heat watered and time restriction relation be:

$y_{{\mathrm{ik}}^d}\left\{\begin{array}{c}=1,{\mathrm{wa}}_{{\mathrm{ik}}^d}\&NotEqual;{\mathrm{wa}}_{i(k^d-1)},\&ForAll;i\&Element;\&lsqb;1,I\&rsqb;,\&ForAll;k^d\&Element;\&lsqb;1,K_i^d\&rsqb;\\ =0,{\mathrm{wa}}_{{\mathrm{ik}}^d}={\mathrm{wa}}_{i(k^d-1)},\&ForAll;i\&Element;\&lsqb;1,I\&rsqb;,\&ForAll;k^d\&Element;\&lsqb;1,K_i^d\&rsqb;\end{array}\right.---\left(20\right)$

${\mathrm{\&tau;}}_s\&le;x_{{\mathrm{ik}}^d}\&le;{\mathrm{\&tau;}}_e,\&ForAll;i\&Element;\&lsqb;1,I\&rsqb;,\&ForAll;k^d\&Element;\&lsqb;2,K_i^d\&rsqb;---\left(21\right)$

${\mathrm{\&tau;}}^{\&prime;}+{\mathrm{\&tau;}}_i^{gap}\&CenterDot;y_{{\mathrm{ik}}^d}\&le;x_{{\mathrm{ik}}^d}\&le;{\mathrm{\&tau;}}^{\&prime;}+({\mathrm{\&tau;}}_e-{\mathrm{\&tau;}}^{\&prime;})\&CenterDot;y_{{\mathrm{ik}}^d},\&ForAll;i\&Element;\&lsqb;1,I\&rsqb;,\&ForAll;k^d\&Element;\&lsqb;1,K_i^d\&rsqb;---\left(22\right)$

(20) formula forces disconnected watering when representing heat discontinuity surface specification difference,

(21) formula represent chosen respectively wait to open water heat casting time and be in the plan phase,

(22) formula represent when wait open water heat and legacy tasks even water time, casting time point takes legacy tasks finish time, τ '= rq_i/ρ·ra_i·rs_i, when non-company waters, take legacy tasks finish time add heat interval to plan the end of term between certain moment,When respectively connecting each other water until blow-on time, this casting time point takes front heat finish time,When non-company waters, take front heat finish time add heat interval to plan the end of term between some time Carve,

3. continuous casting unit as claimed in claim 1 waters heat selection, sequence and the Multipurpose Optimal Method of casting time decision-making, It is characterized in that, coded method described in claim 1 step S4 is:

S31, total heat number K of each casting machine i in adding up pre-scavenger_i, each heat is pressed the casting time that batch plan is predeterminedFirst Rear sequence, and give heat sequence number to successively each stove after sequence, set up each casting machine heat sequence number reference table, and ensure heat The order of sequence number casting time predetermined with batch plan order keeps consistent；

S32, determines selected heat quantitative range, a length of K of stochastic generation_iEach casting machine binary sequence, with heat sequence number reference Table one_to_one corresponding, waters the restriction relation of heat relation according to preliminary election tank furnace in Production Lot Planning time determine selected heat with waiting to open Quantitative range；

S33, the chromogene section of each casting machine is connected and forms a complete chromosome, and randomly generates the initial of setting scale Population.

4. continuous casting unit as claimed in claim 1 waters heat selection, sequence and the Multipurpose Optimal Method of casting time decision-making, It is characterized in that, described coding/decoding method is:

S41, processes illegal chromosome: the illegal chromosome after cross and variation can be divided into two classes: selected wait to open that to water heat total The number heat quantitative range upper limit more than claim 3 calculating and the heat quantitative range lower limit less than claim 3 calculating, right Become 0 by have more constraint upper limit number 1 at random in the former, the latter be will be less than at random to retrain the 0 of lower limit number and become 1, The chromosome not violating constraint is constant；

S42, produces to wait to open according to the heat order from left to right occurring 1 in the chromosome of each casting machine and waters heat sequence, compare stove Heat feature in sequence number reference table and batch plan, produces in conjunction with constraint (20) formula

S43, waters if certain heat treated in blow-on time sequence is disconnected with front heat and has earliest available time, by constraint (19), (21) Formula determinesOtherwise, determine according to constraint (21-22) formula

5. continuous casting unit as claimed in claim 1 waters heat selection, sequence and the Multipurpose Optimal Method of casting time decision-making, It is characterized in that, determine that elite solution diversity method is:

S51, merges parent and child chromosome, dominance relation between definition individuality, gives each individual imparting sequence according to dominance relation Number grade also sorts, and produces the non-dominant individual collections of different sequence number grade, records non-dominant individual amount in each sequence number grade；

S52, according to the order from small to large of non-dominant individual amount, sequence number grade in elite disaggregation capacity, each sequence number grade, really Determine the maximum sequence number that elite disaggregation can accommodate；

S53, calculate in elite disaggregation with the number summation that solves under current non-dominant grade, and judge that the number summation of described solution is No more than elite disaggregation scale, if it is, perform step S54, if it is not, perform step S55；

S54, calculates the individual crowding distance of current non-dominant grade descending, elite solution is pressed crowding distance from big to small Order is sequentially added into elite disaggregation；

S55, calculates the crowding distance under current non-dominant grade and elite solution is added elite disaggregation；

S56, it may be judged whether reach elite disaggregation scale, if reached, then performs step S11 as claimed in claim 1, if Do not reach, then make the sequence number grade of non-domination solution add 1, perform step S53.

6. continuous casting unit as claimed in claim 1 waters heat selection, sequence and the Multipurpose Optimal Method of casting time decision-making The Multipurpose Optimal Method watering heat with time decision-making left by conticaster, it is characterised in that elite solution adds the rule of elite disaggregation For: set elite disaggregation size, press grade sequence number order from small to large wherein and add individuality, until elite disaggregation is filled out Full, in same sequence number grade, when meet sequence number identical two individual time, preferential add the individuality that crowding distance is big, beyond essence The individuality of English disaggregation size is abandoned.

7. continuous casting unit as claimed in claim 1 waters heat selection, sequence and the Multipurpose Optimal Method of casting time decision-making, It is characterized in that, selecting maximum satisfaction scheme and open as continuous casting and water heat time decision method, described fuzzy optimizing method is such as Under:

S71, the proportion ω of each individuality in calculating target function value set_(r,m), ω_(r,m)Represent m-th object function in individual r Value proportion,The minimum of m-th target function value in expression object function value set, maximum respectively:

${\mathrm{\&omega;}}_{(r,m)}=\left\{\begin{array}{cc}1,& f_{(r,m)}\&le;f_m^{\min }\\ \frac{f_{\left(m\right)}^{\max }-f_{(r,m)}}{f_{\left(m\right)}^{\max }-f_{\left(m\right)}^{\min }},& f_{\left(m\right)}^{\min }<f_{{\mathrm{eval}}_{(r,m)}}<f_{\left(m\right)}^{\max }\\ 0,& f_{(r,m)}\&GreaterEqual;f_{\left(m\right)}^{\max }\end{array}\right.---\left(23\right)$

S72, satisfaction ω of all individualities of standardization_r, wherein N is the population scale of elite disaggregation；

${\mathrm{\&omega;}}_r=\frac{\underset{m=1}{\overset{4}{\&Sigma;}}{\mathrm{\&omega;}}_{(r,m)}}{\underset{r=1}{\overset{N}{\&Sigma;}}\underset{m=1}{\overset{4}{\&Sigma;}}{\mathrm{\&omega;}}_{(r,m)}}---\left(24\right)$

S73, takes the individuality of standardization Maximum SatisfactionHeat and time decision scheme z is watered for finally opening_ik、