CN1753008B - Method of optimization hot rolling scaduled sequence - Google Patents

Abstract

A method for optimizing the rolling plan sequence of hot rolling, including two steps of rolling plan sequence adjustment and rolling plan material sequence adjustment, the rolling plan sequence adjustment step adopts simulated annealing algorithm to change the rolling plan sequence; The material adjustment step of the rolling plan adopts a variable neighborhood search algorithm, which includes changing the processing sequence of the main material part in each rolling plan and adding, deleting, and position exchange adjustment of the main material between rolling plans. The method of the invention greatly facilitates the scheduling work of the planners, and liberates the planners from complicated manual planning. At the same time, this method provides an overall optimal solution, which has great practical significance for the optimal design of hot rolling rolling plan.

Description

A kind of method of optimizing hot rolling scaduled sequence

Technical field

Metallurgy industry manufacturing execution system (MES, Manufacturing Execution System) hot rolling production plan field.

Background technology

Hot rolling scaduled is the instruction of indication slab rolling order.Along with the adjustment of continuous casting plan and the enforcement of production, woven rolling scaduled in slab may change break time, need realize rolling scaduled global optimization by the adjustment of rolling scaduled order adjustment and rolling scaduled material for this variation.When plan is adjusted, to consider rolling procedure, take into account many complicated factors such as contractual delivery phase, continuous casting hot rolling linking, subsequent handling requirement again.The artificial result who is difficult to obtain a global optimization that adjusts.

Summary of the invention

At the problems referred to above, the invention provides a kind of method of optimizing hot rolling scaduled sequence with model.

The optimization of rolling scaduled sequence is to make on the basis of considering many complicated factors such as contractual delivery phase, rolling procedure, continuous casting hot rolling linking, subsequent handling requirement, mainly does following consideration on handling:

Hot rolling material: because the restriction of value volume and range of product, hot rolling material does not have adjustable leeway substantially, only adjusts the main body material when adjusting material, does not relate to hot rolling material;

Optimisation strategy: rolling scaduled sequence optimization can be by adjusting rolling scaduled sequence and adjusting two kinds of realizations of material of rolling scaduled, if but adjust the rolling scaduled order and the order of material simultaneously, need the constraint of consideration simultaneously too much, and two kinds of adjustment interact, and are unfavorable for the efficient realization of algorithm.In view of this consideration, when optimizing rolling scaduled sequence, realize in two steps---rolling scaduled order adjustment and rolling scaduled material adjustment.Wherein rolling scaduled order adjustment only changes rolling scaduled precedence, the material that does not change each rolling scaduled inside and comprised; Rolling scaduled material adjustment change each rolling scaduled in main body material order of partly processing and rolling scaduled between the adjustment such as interpolation, deletion, place-exchange of main body material.By rolling scaduled order adjustment and rolling scaduled material adjustment, reach the target that improves the hot charging ratio, improves continuous cold charge and hot charging burst length.

Characteristics according to rolling scaduled order adjustment and rolling scaduled material adjustment problem: the object of processing all is (the rolling scaduled or slab) that disperses, the target of problem is adjusted hot charging ratio of maximization or the burst length that maximizes continuous cold charge/hot charging, and the technological procedure constraint is more.These two problems all belong to the combinatorial optimization problem that has complicated constraint, can set up corresponding integer programming model.

Rolling scaduled order the institute might arrange can Individual (n is a slab quantity), owing to do not break original rolling scaduled in the rolling scaduled order adjustment process, what mainly change after the adjustment is exactly the rolling time of slab, so most technological procedures all needn't be checked in the process of algorithm search, after obtaining a new rolling scaduled sequence, only need the break time and the rolling time of check slab whether to mate.The constraint that problem will satisfy is handled relatively easily, and simulated annealing is adopted in the algorithm for design decision.This algorithm has had the model of comparative maturity, and comparatively successful application is arranged on combinatorial optimization problem.

Rolling scaduled material is adjusted the problem relative complex, comprises in possible adjustment mode: from certain rolling scaduled deletion slab and the slab of deleting inserted change mutually in other rolling scaduled, with the slab position in rolling scaduled and the position adjustment in whole continuous cold charge/hot charging interval etc.Algorithm requires to realize the requirement of these several adjustment modes respectively, and these adjustment modes can be carried out suitable combination.The algorithm for design decision is adopted and is become the neighborhood search algorithm, and this algorithm is realized these adjustment modes by setting the corresponding different adjustment mode of different neighbour structures in the neighborhood search process.Owing to broken original rolling scaduledly in adjusting, become the neighborhood search algorithm and need differentiate in violation of rules and regulations with technology and combine application.

This inventive method is very easy to staff planners' scheduling work, and staff planners are freed from numerous and diverse artificial planning.This method provides the solution of a total optimization simultaneously, and hot rolling scaduled optimal design is had huge realistic meaning.

Further specify the present invention below in conjunction with drawings and Examples.

Description of drawings

Fig. 1 optimizes process flow diagram for rolling scaduled sequence;

Fig. 2 is that rolling scaduled order is adjusted flow process;

Fig. 3 is that rolling scaduled slab is adjusted flow process.

Embodiment

A kind of method of optimizing hot rolling scaduled sequence, realize in two steps: at first being rolling scaduled order adjustment, is the sequencing of least unit plan for adjustment with rolling scaduled, does not relate to the change of the inner material of plan; Carry out rolling scaduled material adjustment then on the basis of rolling scaduled order adjustment, this step is whole to be unit with the slab, only relates to the change of rolling scaduled inside and rolling scaduled 's material.

The overall principle of rolling scaduled sequence optimization is to do limited adjustment on the basis of the rolling scaduled unit of determining, does not do large-scale adjustment.If violate the intermediate solution of process rule, general strategy is abandoned it exactly, and does not do the adjustment at rule in the algorithm search process.Rolling scaduled sequence is optimized process flow diagram as shown in Figure 1.

(1) rolling scaduled order adjustment

Rolling scaduled order adjustment be a complexity be the factorial level (O (n! )) combinatorial optimization problem, can not solve with the method in traversal search space.Find the solution employing simulated annealing, simulated annealing is a kind of intelligent algorithm of comparative maturity, is applicable to the combinatorial optimization problem that the search volume is very big.In order to improve the performance of algorithm, in algorithm, add the data structure of a taboo table, keep several the separating of search recently.

The form of expression of separating of rolling scaduled order adjustment model is a sequence of rolling scaduled number, determines that with rolling scaduled order the corresponding hot charging in back is than weighing the quality of separating.

Model description:

$\max \frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{C}_{\mathrm{fi}}}{n}---\left(1\right)$

$\min \underset{i=1}{\overset{n}{\mathrm{\Σ}}}(t_{\mathrm{ri}}-t_{\mathrm{ci}})---\left(2\right)$

s.t. t _ri≥t _ci+T i∈{1，2，A，n} (3)

The total quantity of n slab wherein;

C _FiThe hot charging of i piece slab than mark (i=1,2 ..., n), 1 is hot charging, 0 is cold charge;

t _RiThe rolling time of i piece slab;

t _CiThe break time of i piece slab;

The T continuous casting is put the time to the biography of hot rolling.

The optimization aim of formula (1), the rolling scaduled order adjustment of formula (2) expression is adjusted hot charging ratio of maximization and the total residence time in the slab storehouse that minimizes all slabs.This is a multiple goal combinatorial optimization problem, and two targets have different priority, judges that in the algorithm operational process quality of feasible solution uses the value of formula (1), formula (2) to come comparison successively.The requirement that the rolling scaduled order adjustment of formula (3) expression must be observed break time and rolling time coupling, the rolling time of slab must be not less than break time and biography is put the time sum.Because it is original rolling scaduled that rolling scaduled order adjustment does not destroy, so need not to consider the technological procedure constraint in the optimizing process.

Solution procedure:

Rolling scaduled order adjustment model is separated a sequence of rolling scaduled number, with rolling scaduled order determine the corresponding hot charging in back than and the slab quality of separating in storehouse total residence time evaluation.

Solution procedure:

1) provides rolling scaduled order initial solution X0, with annealing temperature initial value T0, threshold value Tmin, annealing temperature drop coefficient α, calculate with the corresponding hot charging of X0 than R0 and slab at storehouse total residence time t, set up the empty array T1 of one dimension and show, make optimum solution Xbest=X0 as taboo, annealing temperature T=T0 upgrades the taboo table with T0;

2) 1 X of picked at random in the neighborhood of X0;

3) judge that X whether in T1, if X was relayed to for the 2nd step at T1, if X upgrades T1 with X not in T1, forwards 4 to then);

4) judge the time coupling (formula (3)) of X, if satisfy hot charging that the time coupling requires to calculate X than R and slab in the storehouse total residence time and forward 5 to), directly do not forward 2 to if match);

5) judge whether X is better than X0,, judge whether X is better than Xbest, make Xbest=X and forward 7 to if X is better than Xbest if X is better than X0), directly do not forward 7 to if X is not better than Xbest); If X is not better than X0, forward 6 to);

6) probability of acceptance of computational solution

Generate random number rand ∈ [0,1], if P 〉=rand forwards 7 to), otherwise forward 2 to);

7) make X0=X, R0=R, T=T* α

8) judge whether to reach the algorithm stop condition,, provide Xbest) as finally separating, if not forwarding 2 to if stop.Rolling scaduled order is adjusted flow process as shown in Figure 2.

Illustrate:

Neighbour structure: scope and the relationship between efficiency of separating the neighbour structure setting of X and algorithm search are very big, here neighbour structure is set at arbitrarily selected two set that the rolling scaduled all possible rolling scaduled sequences that its order exchange is formed are formed, and separates that into picked at random two are rolling scaduled to exchange its order for one of picked at random in this neighborhood.

Initial solution: initial solution can adopt heuritic approach to generate, the summation that adds up the break time that each is rolling scaduled, with rolling scaduled according to break time and series arrangement from small to large, as the initial solution of rolling scaduled order adjustment.What also rolling scaduled establishment can be formed is rolling scaduled as initial solution.

The acceptance probability of separating: though the validity of simulated annealing be embodied in algorithm search to feasible solution be inferior to current separating, algorithm will be accepted it according to certain probability and separate as current.This mechanism makes simulated annealing have to jump out the ability of locally optimal solution.Acceptance probability is relevant with annealing temperature with the hot charging ratio of separating in this algorithm, along with this probability of running time of algorithm reduces gradually.

(2) rolling scaduled material order is adjusted

The material adjustment is to carry out under the prerequisite that rolling scaduled order adjustment has been finished, and is that unit adjusts order and position with the slab in rolling scaduled, mainly is rolling scaduled slab position adjustment.Target is to reduce cold and hot base to load in mixture number of times as far as possible, improves the quantity of continuous cold base, Continuous Heat base.

The goal-setting of material adjustment model is the slab number that minimizes continuum number and the maximum continuum of maximization; The constraint condition that will satisfy in the adjustment process is that technological procedure constraint and rolling time---mate break time.

According to the continuum indefinite length, material should allow the characteristics of monolithic slab or whole continuum shift position when adjusting, need to set the adjustment effect that reaches different by different neighborhoods in the search procedure of algorithm.Adopt to become the neighborhood search algorithm, this algorithm is set multiple neighbour structure, is used for the different material of respective material adjustment and moves or the exchange state.

The form of expression of separating of rolling scaduled material adjustment model is carries out adjusted rolling scaduled sequence, the quality of separating is differentiated according to the order of continuum number, maximum continuum slab number relatively, the number of more maximum continuum again under the two situation about all equating.

Adopt three kinds of neighbour structures: single slab inserts, slab exchanges, the continuum exchange.

Model description:

$\max \frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{C}_{\mathrm{fi}}}{n}---\left(4\right)$

$\min \underset{i=1}{\overset{n}{\mathrm{\Σ}}}(t_{\mathrm{ri}}-t_{\mathrm{ci}})---\left(5\right)$

s.t. t _ri≥t _ci+T (6)

Adjacent slab maximum rolling width jump value constraint (7)

Adjacent slab maximum rolling thickness jump value constraint (8)

Adjacent slab highest hardness jump value constraint (9)

$\begin{array}{ccc}s.t.& \underset{k=1}{\overset{m}{\mathrm{\Σ}}}{y}_{\mathrm{ik}}=1& i=\mathrm{1,2,3},\mathrm{\Λ},n---\left(10\right)\end{array}$

$\begin{array}{ccc}\underset{j=1}{\overset{n_{\mathrm{zk}}}{\mathrm{\Σ}}}{z}_{\mathrm{ijk}}{q}_j{y}_{\mathrm{jk}}\≤A& k=\mathrm{1,2,3},...,m,& i=\mathrm{1,2},...,n_{\mathrm{zk}}---\left(11\right)\end{array}$

$\begin{array}{ccc}{\mathrm{BMin}}_j\≤\underset{j=1}{\overset{n_{\mathrm{zk}}}{\mathrm{\Σ}}}{B}_{\mathrm{ijk}}{q}_j{y}_{\mathrm{jk}}\≤{\mathrm{BMax}}_j& k=\mathrm{1,2,3},...,m,& i=\mathrm{1,2},...,n_{\mathrm{zk}}---\left(12\right)\end{array}$

$\begin{array}{ccc}{\mathrm{GMin}}_j\≤\underset{j=1}{\overset{n_{\mathrm{zk}}}{\mathrm{\Σ}}}{G_{\mathrm{ijk}}q}_j{y}_{\mathrm{jk}}\≤{\mathrm{GMax}}_j& k=\mathrm{1,2,3},...,m,& i=\mathrm{1,2},...,n_{\mathrm{zk}}---\left(13\right)\end{array}$

$\begin{array}{ccc}{\mathrm{TMin}}_j\≤\underset{j=1}{\overset{n_{\mathrm{zk}}}{\mathrm{\Σ}}}{T_{\mathrm{ijk}}q}_j{y}_{\mathrm{jk}}\≤{\mathrm{TMax}}_j& k=\mathrm{1,2,3},...,m,& i=\mathrm{1,2},...,n_{\mathrm{zk}}---\left(14\right)\end{array}$

$\begin{array}{ccc}{\mathrm{LMin}}_j\≤\underset{j=1}{\overset{n_{\mathrm{zk}}}{\mathrm{\Σ}}}{L_{\mathrm{ijk}}w}_j{y}_{\mathrm{jk}}\≤{\mathrm{LMax}}_j& k=\mathrm{1,2,3},...,m,& i=\mathrm{1,2},...,n_{\mathrm{zk}}---\left(15\right)\end{array}$

The total quantity of n slab wherein;

C _FiThe hot charging of i piece slab than mark (i=1,2 ..., n) 1 is hot charging, 0 is cold charge;

t _RiThe rolling time of i piece slab;

t _CiThe break time of i piece slab;

The T continuous casting is put the time to the biography of hot rolling.

The rolling scaduled number of m

n _ZkMain body material slab number in the rolling scaduled k;

q _jThe mill length of slab j;

w _jThe weight of slab j;

The rolling length constraint of A with the width slab;

BMin _j, BMax _jThe length constraint of rolling slab with surface grade

GMin _j, GMax _jThe length constraint of rolling slab with steel grade

TMin _j, TMax _jThe length constraint of rolling slab with cold and hot base type

LMin _j, LMax _jRolling weight constraints with the slab that flows to

The optimization aim of formula (4), the rolling scaduled material adjustment of formula (5) expression is respectively the adjusted hot charging ratio of maximization and minimizes the total residence time in the slab storehouse of all slabs.This is a combinatorial optimization problem, and two targets have different priority, judges that in the algorithm operational process quality of feasible solution uses the value of formula (4), formula (5) to come comparison successively.The rolling scaduled material adjustment of formula (6) expression also needs to satisfy the coupling of break time and rolling time, and formula (7)-(15) the rolling scaduled material adjustment of expression need be satisfied the constraint of all production technology rules.

Solution procedure

The form of expression of separating of rolling scaduled material adjustment model is rolling scaduled sequence, and the quality differentiation of separating is always compared in the order of the storehouse residence time (formula (5)) than (formula (4)), slab according to hot charging.

Adopt three kinds of neighbour structures: single slab inserts, slab exchanges, the continuum exchange.

Solution procedure

1) initialization procedure: the Nk that determines in the algorithm to use (k=1,2 ..., kmax) plant neighbour structure; Determine initial solution x; Set up historical optimum solution P, upgrade P with x;

2) make k=1, the iterative process of beginning algorithm;

3) 1 x ' of picked at random in the k class neighborhood of x;

4) judge that whether x ' satisfies the rules constraint, does not forward 3 to if do not satisfy), continue execution in step 5 if satisfy);

5) be better than x if satisfy x ', make x=x ' forward 3 then to) with x ' renewal P;

6) judge whether to satisfy k=kmax,, then make k=1, forward 3 to) if k has reached maximal value; If reaching maximal value, k do not forward 7 to);

7) judge whether to satisfy stop condition, stop, export historical optimum solution P, do not make k=k+1 forward 3 to) if do not satisfy if satisfy algorithm.

Rolling scaduled slab is adjusted flow process as shown in Figure 3.

Illustrate:

Neighborhood is set: adopt three kinds of neighbour structures---slab insertion, slab place-exchange and continuum exchange.Wherein the slab insertion is to select a certain block of slab to insert assigned address, and the slab place-exchange is the location swap with the two boards base of appointment, and the continuum exchange is the location swap with two continuums of appointment.

Rules constraint qualification: for the adjusted rolling scaduled rules verification (mostly be most two rolling scaduled) of doing, check the situation that satisfies of rule successively, violate situation and just be judged to be and do not satisfy rule constrain as long as rules occur according to formula (7)-(15).In algorithm is realized in order to improve the operational efficiency of algorithm, the rules check divides two parts to carry out, and retrains the rules (hop limit of rolling width, rolling thickness, rolling hardness, temperature etc.) that must observe between the rolling scaduled middle slab and is carrying out mobile (insertion, exchange, the exchange of polylith slab) preceding testing of slab; Carry out the slab move operation again under the prerequisite of the constraint more than satisfying, carry out other technological procedure checks then.

Upgrade historical optimum solution: after obtaining an intermediate solution in the algorithm, itself and historical optimum solution are made comparisons, keep a more excellent historical optimum solution of conduct.

The sign continuum: judge successively each rolling scaduled in the cold and hot dress state of each slab, the slab of cold charge or hot charging distributes same unique continuum identification number continuously, identifies the sequence number of slab in continuum simultaneously.

Claims (3)

1. A method for optimizing the sequence of hot-rolled rolling plans, comprising two major steps of rolling plan sequence adjustment and rolling plan material sequence adjustment, the rolling plan sequence adjustment step adopts simulated annealing algorithm to change the sequence of rolling plans Sequence: The rolling plan material sequence adjustment step adopts a variable neighborhood search algorithm, which includes changing the processing order of the main material part in each rolling plan and the addition, deletion, and position exchange adjustment of the main material between the rolling plans; The step of adjusting the sequence of the rolling plan includes establishing a first model whose solution is a sequence of rolling plan numbers, and evaluates the quality of the solution by using the corresponding thermal charging ratio and the total residence time of the slab after the rolling plan sequence is determined. In the simulated annealing algorithm of the rolling plan sequence adjustment step, a data structure of a taboo table is added to keep several solutions recently searched; the rolling plan material sequence adjustment step includes establishing a second model, based on the rolling plan Adjust the order and position of the slab as a unit; use the variable neighborhood search algorithm to set three neighborhood structures: single slab insertion, slab exchange, and continuous interval exchange; The first model includes the following formulas: $\mathrm{<span\; class="notranslate">\; max</span>}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; the\; fi</span>}}}{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$ $\mathrm{<span\; class="notranslate">\; min</span>}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; the\; ri</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; ci</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$ st t _ri ≥ t _ci + T i ∈ {1, 2, ..., n} (3) where n is the total number of slabs; C _fi hot charging ratio mark of the i-th slab, where i=1, 2,...,n; 1 is hot charging, 0 is cold charging;

t _ri is the rolling time of the i-th slab; t _{ci is} the cutting time of the i-th slab; T The transfer time from continuous casting to hot rolling; Formulas (1) and (2) indicate that the optimization goal of rolling plan sequence adjustment is to maximize the adjusted hot charging ratio and minimize the total residence time of all slabs in the slab storehouse; The quality of the feasible solution is compared with the values of formula (1) and formula (2) in turn; formula (3) indicates that the adjustment of the rolling plan sequence must comply with the requirement of matching the cutting time and rolling time, that is, the rolling time of the slab must be Not less than the sum of cut-off time and transmission time; The specific solution steps are: 1) Given the initial solution X0 of the rolling plan sequence, the initial value T0 of the annealing temperature, the threshold value Tmin, and the annealing temperature drop coefficient α, calculate the thermal charging ratio R0 corresponding to X0 and the total residence time t of the slab in the warehouse, and establish a The empty-dimensional array T1 is used as a taboo table, so that the optimal solution Xbest=X0, the annealing temperature T=T0, and update the taboo table with T0; 2) Randomly select a point X in the neighborhood of X0; 3) Judging whether X is in T1, if X is in T1, go to step 2), if X is not in T1, update T1 with X, and then go to 4); 4) Apply the formula (3) to judge the time matching of X, if the time matching requirement is met, calculate the thermal charging ratio R of X and the total residence time of the slab in the warehouse and go to 5), if it does not match, go to 2); 5) judge whether X is better than X0, if X is better than X0, judge whether X is better than Xbest, if X is better than Xbest then make Xbest=X and go to 7), if X is not better than Xbest directly go to 7); If X is not better than X0, go to 6); 6) Calculate the acceptance probability of the solution

Generate a random number rand∈[0,1], if P≥rand, go to 7), otherwise go to 2); 7) Let X0=X, R0=R, T=T*α; 8) Judging whether the stop condition of the algorithm is reached, if so, give Xbest as the final solution, if not, go to 2); the neighborhood structure of the solution X is set to be formed by arbitrarily selecting two rolling plans and exchanging their order A set of all possible rolling plan sequences, a solution randomly selected in the neighborhood is two randomly selected rolling plans to exchange their order; the initial solution is generated by a heuristic algorithm, and the cut-off of each rolling plan Time accumulation and summation, the rolling plan is arranged in ascending order of the sum of cut-off time, as the initial solution for the adjustment of the rolling plan sequence, or the rolling plan formed by the rolling plan is used as the initial solution; The second model includes the following formula: $\mathrm{<span\; class="notranslate">\; max</span>}\frac{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; the\; fi</span>}}}{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$ $\mathrm{<span\; class="notranslate">\; min</span>}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; the\; ri</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; ci</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$ st t _ri ≥t _ci +T (6) Constraints on jumping value of maximum rolling width of adjacent slabs (7) Constraints on the maximum rolling thickness jump value of adjacent slabs (8) Constraints on the maximum hardness jump value of adjacent slabs (9) st $\underset{k=1}{\overset{m}{\mathrm{\&Sigma;}}}{\mathrm{the\; y}}_{\mathrm{ik}}=1$ i=1, 2, 3, ..., n (10) $\underset{j=1}{\overset{{\mathrm{no}}_{\mathrm{zk}}}{\mathrm{\&Sigma;}}}{z}_{\mathrm{ijk}}{q}_j{\mathrm{the\; y}}_{\mathrm{jk}}\&le;A$ k=1, 2, 3,..., m, i=1, 2,..., n _zk (11) ${\mathrm{BMin}}_j\&le;\underset{j=1}{\overset{{\mathrm{no}}_{\mathrm{zk}}}{\mathrm{\&Sigma;}}}{B}_{\mathrm{ijk}}{q}_j{\mathrm{the\; y}}_{\mathrm{jk}}\&le;{\mathrm{BMax}}_j$ k=1, 2, 3,..., m, i=1, 2,..., n _zk (12) ${\mathrm{GMin}}_j\&le;\underset{j=1}{\overset{{\mathrm{no}}_{\mathrm{zk}}}{\mathrm{\&Sigma;}}}{G}_{\mathrm{ijk}}{q}_j{\mathrm{the\; y}}_{\mathrm{jk}}\&le;{\mathrm{GMax}}_j$ k=1, 2, 3, . . . , m, i=1, 2, . . . , n _zk (13) ${\mathrm{TMin}}_j\&le;\underset{j=1}{\overset{{\mathrm{no}}_{\mathrm{zk}}}{\mathrm{\&Sigma;}}}{T}_{\mathrm{ijk}}{q}_j{\mathrm{the\; y}}_{\mathrm{jk}}\&le;{\mathrm{TMax}}_j$ k=1, 2, 3, . . . , m, i=1, 2, . . . , n _zk (14) ${\mathrm{LMin}}_j\&le;\underset{j=1}{\overset{{\mathrm{no}}_{\mathrm{zk}}}{\mathrm{\&Sigma;}}}{L}_{\mathrm{ijk}}{w}_j{\mathrm{the\; y}}_{\mathrm{jk}}\&le;{\mathrm{LMax}}_j$ k=1, 2, 3,..., m, i=1, 2,..., n _zk (15) where n is the total number of slabs; C _fi hot charging ratio mark of the i-th slab, where i=1, 2,...,n; 1 is hot charging, 0 is cold charging; t _ri is the rolling time of the i-th slab; t _{ci is} the cutting time of the i-th slab; T The transfer time from continuous casting to hot rolling; m number of rolling plans;

n _zk is the number of main material slabs in rolling plan k; q _j is the rolling length of slab j; w _j weight of slab j; A length constraint of rolling slabs of the same width; Length constraints of BMin _j and BMax _j rolling slabs with the same surface grade Length constraint of GMin _j and GMax _j rolling slabs of the same steel type Length constraint of TMin _j , TMax _j rolling slabs of the same type as cold and hot slabs Weight constraints of LMin _j and LMax _j rolling slabs with the same flow direction Equation (4) and Equation (5) indicate that the optimization objectives of material sequence adjustment in the rolling plan are to maximize the adjusted hot charging ratio and minimize the total residence time of all slabs in the slab warehouse; the two objectives have Different priorities, the quality of the feasible solution is judged during the operation of the algorithm and compared with the values of formula (4) and formula (5); formula (6) indicates that the material adjustment of the rolling plan also needs to meet the cutting time and rolling time The matching of , the formulas (7)-(15) indicate that the adjustment of the material sequence of the rolling plan needs to meet the constraints of all production process regulations; The specific solution steps are: a) Initialization process: determine the Nk kinds of neighborhood structures that will be used in the algorithm, where k=1, 2, ..., kmax; determine the initial solution x; establish the historical optimal solution P, and update P with x; B) make k=1, start the iterative process of algorithm; c) Randomly select a point x' in the kth neighborhood of x; d) Judging whether x' satisfies the constraints of the production process regulations, if not satisfied, go to c), if satisfied, continue to execute step e); e) If it is satisfied that x' is better than x, set x=x' to update P with x' and then go to c); otherwise, proceed to step f); f) judging whether k=kmax is satisfied, if k has reached the maximum value, then make k=1, and go to c); if k does not reach the maximum value, go to g); g) Judging whether the stop condition is satisfied, if the algorithm is satisfied, the algorithm is terminated, and the historical optimal solution P is output, if not, set k=k+1 and go to c). 2. A method for optimizing the sequence of hot-rolling rolling plans according to claim 1, characterized in that: for the adjusted rolling plan to check the production process regulations, check in turn according to formulas (7)-(15) Satisfaction of the production process regulations, as long as there is a violation of the production process regulations, it will be judged as not meeting the constraints of the production process regulations. 3. A method for optimizing the sequence of hot-rolling rolling plans according to claim 2, characterized in that: the verification of the production process specification is carried out in two parts, and the rolling plan that must be complied with between slabs in the rolling plan is restricted. The rolling width, rolling thickness, and rolling hardness are inspected before the slab is moved; the slab is moved under the premise of satisfying the above constraints, and then other process specification inspections are performed.

Images