CN104281134B - System and method for optimizing multiple production indexes in sorting process of raw ore based on man-machine interaction - Google Patents

Abstract

The invention discloses a system and method for optimizing multiple production indexes in the sorting process of raw ore based on man-machine interaction, and belongs to the technical field of ore-dressing production processes. The system comprises a data acquisition module, a parameter setting and modifying unit, an optimization algorithm module library, an optimization model module, an optimization result output unit and a multi-production index optimization system management unit. The method includes the steps that the optimization algorithm module library is searched for a needed optimization algorithm first according to an optimization model; if the algorithm is not found out, the needed optimization algorithm is packaged into a new optimization algorithm module, and if the algorithm is found out, a needed parameter is acquired; an optimization scheme is established; the optimization scheme is operated; a set of production index optimization values are acquired, and meanwhile corresponding raw ore sorting configuration quantity optimal values are issued to an under layer ore dressing production system; if a production index actual value is placed within the production index target value range, the optimization scheme is operated continuously, and if not, the parameter acquiring step is executed again. By the adoption of the system and method, the purposes for saving energy, reducing consumption, lowering production cost and improving economic benefits can be achieved.

Description

Raw ore based on man-machine interaction sorts the many production targets of process and optimizes system and method

Technical field

The invention belongs to dressing Production Process technical field, be specifically related to a kind of raw ore based on man-machine interaction and sort process Many production targets optimize system and method.

Background technology

Along with the development of global economy, the demand of mineral resources is also being continued to increase by various countries, but due to ore deposit Producing resource and belong to non-renewable resources, As time goes on, owing to mining activity is continuously increased, Resource management is in the world On all constantly declining.Under this form, the most effectively sort the usage amount of various raw ore to realize the optimization of ore dressing Journey, becomes the problem that ore dressing plant is in the urgent need to address.

The main task of mineral processing production index optimization decision-making is to optimize the resources of production (i.e. ore resource configuration), to reach pre- The productive target of phase.In these mineral processing production targets, relate to five main production targets, i.e. concentrate grade, entirely select ratio, essence Mineral products amount, metal recovery rate and unit concentrate production cost.Ideally, it is desirable to this five indices all reached in the same time To optimum.But, several conflicting, such as owing to these five indexs having: concentrate grade and gold can not be increased simultaneously Belong to the response rate, excessively pursue metal recovery rate and concentrate grade will be caused to decline；Raw ore selects higher than low then price entirely, pursues height and receives Benefit, may result in high impurity and low-grade etc., and the mineral processing production index in existing ore dressing plant major part is based on manual decision, Preference information amplifies, and causes error relatively big, so production schedule department can not allow all of index reach to optimize, therefore, for Solve these problems, design and develop the many production targets of ore dressing process and optimize that system is quick to Mineral Processing Enterprises, optimally formulate life Produce plan, improve production efficiency, improve production target Decision Quality and have great importance.

But from the point of view of sorting process many production targets optimization systematic research present situation according to domestic and international raw ore, mineral processing in China is looked forward to The thinking that the decision-making of production management department of industry formulates the production schedule is relatively backward, a lot of in the case of be all artificial formulation, additionally The field apparatus of a lot of Mineral Processing Enterprises is more outmoded, lacks the information management apparatus in modern times, the most a lot of many production targets optimizations Decision method can not be applied among scene well, is mainly manifested in (1) and is difficult to the optimization of the production schedule: whole production The formulation process of plan is all accomplished manually by production schedule personnel, and workload is very big, and plan link is comparatively laborious, Crucially production schedule personnel are difficult to take into account all of production target, even cause indivedual production target poor, it is difficult to Realize the optimization of the whole production schedule, it is more difficult to the maximization generated profit.(2) information can not be each production department of Mineral Processing Enterprises Between Men the most alternately: most domestic ore dressing plant uses traditional hierarchical structure enterprise schema, and the production schedule is pressed from head factory According to level, being issued to each subsidiary factory layer by layer, each subsidiary factory is assigned to each production division, each production line, each operation area again, then under Reaching to each class, group, finally arrive the operator at scene, the feedback of information is communicated up the most layer by layer.Due to each department Between mutual independence and the link assigned of the production schedule more so that company information has very when transmitting between each department Big delay, so that the production schedule can not be assigned to the operator of a line rapidly from production schedule department, instructs work People produces；The on-the-spot condition of production can not feed back to decision-making of production management department rapidly, and subsidiary production plan department revises Improve the new production schedule.Accordingly, because information can not inter-sectional interact at each effectively, whole ore dressing can be caused raw Product process efficiency declines.(3) lack a kind of general software and be specifically designed for the production target optimization process of Mineral Processing Enterprises.

Summary of the invention

The shortcomings and deficiencies existed for prior art, the present invention provides a kind of raw ore based on man-machine interaction to sort process Many production targets optimize system and method.

Technical scheme:

A kind of raw ore based on man-machine interaction sorts the many production targets of process and optimizes system, including: production target data acquisition Collection module, parameter set and amendment unit, many production targets optimization algorithm module storehouse, many production targets Optimized model module, many Production target optimum results output unit and many production targets optimize System Management Unit；

Described production target data acquisition module is used for by the OPC communication mechanism with lower floor's mineral processing production system foundation, Gather many production targets actual value of lower floor's mineral processing production system and show；Described many production targets, including: concentrate product Position, entirely select ratio, metal recovery rate, unit Concentrate cost and concentrate yield；

Described parameter sets and amendment unit, for being set the parameter needed for many production targets prioritization scheme and repairing Change, including: production target desired value module, boundary constraint module, optimized algorithm inner parameter module；These modules all pass through phase The human-computer interaction interface answered carries out man-machine interaction；Described parameter, including: many production targets desired value, edge-restraint condition and life Produce index optimization algorithm inner parameter；

Described production target desired value module, is used for the desired value to each production target on corresponding human-computer interaction interface Scope is set and revises；The target range of described production target, including: concentrate grade bound, entirely select the upper of ratio Limit, the lower limit of metal recovery rate, the upper limit of unit Concentrate cost and the bound of concentrate yield；

Described boundary constraint module, for sorting border that process relates to about on corresponding human-computer interaction interface to raw ore Bundle condition is set and revises；Described boundary constraint, including: production target target range, capacity of equipment, tailings grade, Available energy resources, available ore resource；Wherein production target target range is not set in this module and revises, and It is by calling the acquisition of production target desired value module.

Described optimized algorithm inner parameter module, refers to for the production designed user on corresponding human-computer interaction interface The inner parameter of mark optimized algorithm is configured and revises；Described inner parameter, the parameter needed for algorithm itself；

Described many production targets optimization algorithm module storehouse, is constituted for multiple different many production targets optimization algorithm module Set；

Described many production targets optimization algorithm module, for many production targets optimized algorithm of determining of encapsulation, will every kind Many production targets optimized algorithm is encapsulated as production target optimization algorithm module more than；

Described many production targets Optimized model module, is used for loading many production targets Optimized model；

Described many production targets optimum results output unit includes: optimum results output module, production target monitor in real time Module, raw ore sort configuration amount and issue module；

Described optimum results output module is used for output in real time and display by obtaining after the optimization of many production targets prioritization scheme To raw ore sort configuration amount optimal value and production target optimal value；Described raw ore sorts configuration amount and refers to various ore resource Consumption；

Described production target real-time monitoring module is for showing the real-time of production target actual value and production target optimal value Variation tendency；

Described raw ore sorts configuration amount and issues module for by optimum results output module and lower floor's mineral processing production system The OPC communication mechanism set up, sorts corresponding raw ore configuration amount optimal value and is issued to lower floor's mineral processing production system；

Described production target optimizes System Management Unit, including: subscriber information management module, Password Management module, log in/ Exit management module and help module；Described subscriber information management module is used for managing user information, including: rights management, use Family management and log management；Described Password Management module is used for managing user cipher；

Described log in/exit module: be used for managing user and log in/exit process；

Described help module is for according to user's needs, and the system that provides the user uses help information.

A kind of raw ore based on man-machine interaction sorts process many production targets optimization method, described in employing based on man-machine friendship Mutual raw ore sorts the many production targets of process and optimizes system realization, and it comprises the steps:

Step 1: determine many production targets Optimized model, according to many production targets Optimized model, optimizes in many production targets In algoritic module storehouse, the many production targets optimized algorithm needed for searching according to algorithm parameter configuration, if not finding, then perform step 2, if finding, then perform step 3；

Step 2: required many production targets optimized algorithm is packaged into new many production targets optimization algorithm module, and handle This new many production targets optimization algorithm module is added in many production targets optimization algorithm module storehouse, and performs step 1；

Step 2.1: by storing path, sets up many production targets optimized algorithm to be packaged and calculates with the optimization of many production targets Corresponding relation between method module library；

Step 2.2: be provided for reading the algorithm parameter configuration interface of different language, it includes the input and output ginseng of algorithm Number, the title of algorithm；

Step 2.3: configure interface according to the algorithm parameter arranged and algorithm is preserved at most production target optimization algorithm module In storehouse；

Step 3: the parameter needed for acquisition；

Respectively in production target desired value module, boundary constraint module, the man-machine interaction of optimized algorithm inner parameter module On interface, on-line setup or revise corresponding parameter, or, from data base, obtain corresponding parameter；

Step 4: build many production targets prioritization scheme；

Call inside production target data acquisition module, production target desired value module, boundary constraint module, optimized algorithm Parameter module, optimum results output module, raw ore sort configuration amount and issue module, production target real-time monitoring module, produce more Index optimization algoritic module, builds many production targets prioritization scheme；

Step 5: many production targets prioritization scheme that operating procedure 4 builds, then obtain at optimum results output module simultaneously Raw ore sorts configuration amount optimal value and production target optimal value, obtains production target optimal value in production target real-time monitoring module Real-time change tendency information with production target actual value；

Step 6: from the production target optimal value that optimum results output module shows, according to the real-time change of production target value Change tendency information, decisionmaker's preference information and be actually needed, selecting one group of production target optimal value, meanwhile, this group is produced and refers to The raw ore that mark optimal value is corresponding sorts configuration amount optimal value and is issued to lower floor's mineral processing production system；

Step 7: judge that whether production target actual value that production target data acquisition module shows is in production target target In the range of value, it is then to perform step 6, no, then perform step 3.

Beneficial effect: utilize the raw ore based on man-machine interaction of the present invention to sort the many production targets of process and optimize system and side Method, user can pass through grassroot project, according to the self-demand of user, many production targets optimized algorithm is encapsulated into software platform, The raw ore building personalization sorts process many production targets optimization system, reaches energy-saving and cost-reducing, reduces production cost, improves economy The purpose of benefit.Therefore this invention has the advantage that

1) convenient algorithm packaging function: user can as required, self-defined many production targets optimized algorithm, and passes through Native system is packaged into many production targets optimization algorithm module of correspondence, and can configure module graphical interfaces, then by this new envelope Many production targets optimization algorithm module of dress is added in many production targets optimization algorithm module storehouse, and can be to module and module Modify maintenance in the many production targets optimization algorithm module storehouse constituted.

2) many production targets prioritization scheme systematic function efficiently: user can according to oneself needing easily to from Oneself many production targets prioritization scheme carries out modular arrangements flexibly, the operation order of setting module and cycle of operation, finally Many production targets prioritization scheme that formation can run.

3) tables of data and trend monitoring: user can check production respectively by the data carrying out being grouped by functional category Target goals, boundary condition, the concrete data of setting value；By trendgram can check easily production target historical trend and Real-time tendency.

4) OPC communication: by OPC communication function, can conveniently realize raw ore sort configuration amount assign and production refers to Uploading of mark actual value, it is simple to management strategy department adjusts the production schedule in time.

Accompanying drawing explanation

Fig. 1 is that the raw ore based on man-machine interaction of one embodiment of the present invention sorts process many production targets optimization system Structural representation；

Fig. 2 is that the raw ore of one embodiment of the present invention sorts process many production targets optimization method flow chart；

Fig. 3 is the structural representation of many production targets prioritization scheme of one embodiment of the present invention；

Fig. 4 is the production target optimal value real-time change trend with production target actual value of one embodiment of the present invention Figure；

Fig. 5 is that the production target optimized algorithm non-dominated ranking based on gradient information of one embodiment of the present invention is lost Propagation algorithm flow chart.

Detailed description of the invention

Below in conjunction with the accompanying drawings the detailed description of the invention of the present invention is described further.

The raw ore based on man-machine interaction of present embodiment sorts the many production targets of process and optimizes system, as it is shown in figure 1, its Including: production target data acquisition module, parameter set and revise unit, many production targets optimization algorithm module storehouse, produce more Index optimization model module, many production targets optimum results output unit and many production targets optimize System Management Unit；This reality Execute and mode uses building of C# language completion system, use WPF language finishing man-machine interaction interface；

Described production target data acquisition module is for by OPC (the OLE for lower floor's mineral processing production system foundation Process Control) communication mechanism, gather many production targets actual value of lower floor's mineral processing production system and show；Institute State many production targets, including: concentrate grade, entirely select ratio, metal recovery rate, unit Concentrate cost and concentrate yield；

Described parameter sets and amendment unit, for being set the parameter needed for many production targets prioritization scheme and repairing Change, including: production target desired value module, boundary constraint module, optimized algorithm inner parameter module；These modules all pass through phase The human-computer interaction interface answered carries out man-machine interaction；Described parameter, including: many production targets desired value, edge-restraint condition and life Produce index optimization algorithm inner parameter；

Described production target desired value module, is used for the desired value to each production target on corresponding human-computer interaction interface Scope is set and revises；The target range of described mineral processing production index, including: concentrate grade bound, entirely select ratio The upper limit, the lower limit of metal recovery rate, the upper limit of unit Concentrate cost and the bound of concentrate yield；

Described boundary constraint module, for sorting border that process relates to about on corresponding human-computer interaction interface to raw ore Bundle condition is set and revises；Described boundary constraint, including: production target target range, capacity of equipment, tailings grade, Available energy resources, available ore resource；Described equipment/process capability constraint, mainly includes (monthly/Sunday) main equipment Ability, treating capacity, mistake when treating capacity, magnetic separation board when treating capacity, weak magnetic grinding machine platform when treating capacity, strong magnetic grinding machine platform during shaft furnace platform Treating capacity etc. during filter platform；Described production run time constraint condition, mainly includes that shaft furnace, strong magnetic grinding machine, weak magnetic grinding machine etc. are each Kind equipment runs the time, produces each series operation times such as series；Described optimized algorithm inner parameter module, for accordingly On human-computer interaction interface, the inner parameter to the production target optimized algorithm of user's design is configured and revises；Described internal ginseng Number, the parameter needed for algorithm itself, such as decision variable number, object function number, population number, iterations etc.；

In dressing Production Process, production schedule department can be in conjunction with the actual feelings of production target optimum results and produced on-site Condition, by the control parameter that corresponding human-computer interaction interface online modification is relevant, thus adjusts the condition of production at scene in time.

Described many production targets optimization algorithm module storehouse, is constituted for multiple different many production targets optimization algorithm module Set；

Described many production targets optimization algorithm module, for many production targets optimized algorithm of determining of encapsulation, will every kind Many production targets optimized algorithm is encapsulated as production target optimization algorithm module more than；

Described many production targets Optimized model module, is used for loading many production targets Optimized model；Present embodiment is many Production target Optimized model: because Mineral Processing Enterprises production target optimization aim mainly has concentrate yield, concentrate grade, entirely select ratio, Metal recovery rate and unit cost, wherein under meeting concentrate demand and constraints, maximum in the target zone of yield Change concentrate yield；Concentrate grade is maximized in grade target interval；Under the constraints of the ratio of concentration upper limit, reduce choosing as far as possible Ore deposit ratio, reduces the raw ore consumption of unit product；Ore dressing plant is maximized under the constraint of the metal recovery rate lower limit of ore dressing plant setting Metal recovery rate；For profit-push, under the constraint of the given unit cost upper limit, maximize Concentrate cost；Therefore, this reality The production target optimization object function executed in the production target Optimized model of mode includes:

1. maximization concentrate yield:

$MaxQ\left(\stackrel{\→}{X}\right)=Max\left(\underset{i=1}{\overset{I}{\Σ}}\right(1-u_i)x_i/k_{1,i}+\underset{i=1}{\overset{I}{\Σ}}{u}_i{x}_i/k_{2,i})---\left(1\right)$

In formula,For concentrate yield (ten thousand tons)；Decision variableIt is made up of various ore handling capacities, i.e.x_iIt is the treating capacity (ton) of i-th kind of raw ore, i=1,2 ..., I；u_iFor the block fracture slope of raw ore i, The most isolated lump ore amount ratio to raw ore amount；k_1,iRatio of concentration for the fine ore that raw ore i screens out；k_2,iSieve for raw ore i The ratio of concentration of the lump ore gone out；(1-u_i)x_iRepresent that i-th kind of raw ore is through sieving isolated fine ore amount；u_ix_iRepresent i-th kind of raw ore Through sieving isolated lump ore amount；(1-u_i)x_i/k_1,iRepresent concentrate produced by fine ore；u_ix_i/k_2,iRepresent what lump ore produced Concentrate；

2. maximization concentrate grade:

$Max\mathrm{\β}\left(\stackrel{\→}{X}\right)=Max\frac{\underset{i=1}{\overset{I}{\Σ}}(1-u_i){\mathrm{\β}}_{1.i}{x}_i/k_{1,i}+\underset{i=1}{\overset{I}{\Σ}}{u}_i{\mathrm{\β}}_{2.i}{x}_i/k_{2,i}}{\underset{i=1}{\overset{I}{\Σ}}(1-u_i)x_i/k_{1,i}+\underset{i=1}{\overset{I}{\Σ}}{u}_i{x}_i/k_{2,i}}---\left(2\right)$

In formula,For concentrate grade (%)；β_1.iHigh intensity magnetic mineral grade for raw ore i；β_2.iWeak magnetic essence for raw ore i Ore deposit grade；Because concentrate grade represents the tenor in concentrate product, then denominator is the yield of concentrate, and molecule is high intensity magnetic mineral Summation with amount of metal contained in inferior fine magnetite concentrate.

3. minimize and entirely select ratio:

$MinK\left(\stackrel{\→}{X}\right)=Min\frac{\underset{i=1}{\overset{I}{\Σ}}{x}_i}{\underset{i=1}{\overset{I}{\Σ}}(1-u_i)x_i/k_{1,i}+\underset{i=1}{\overset{I}{\Σ}}{u}_i{x}_i/k_{2,i}}---\left(3\right)$

In formula,For full choosing ratio (again), full choosing is than the ratio equal to total raw ore amount with total concentrate yield.

4. maximization metal recovery rate:

$Max\mathrm{\ϵ}\left(\stackrel{\→}{X}\right)=Max\frac{\underset{i=1}{\overset{I}{\Σ}}(1-u_i){\mathrm{\β}}_{1.i}{x}_i/k_{1,i}+\underset{i=1}{\overset{I}{\Σ}}{u}_i{\mathrm{\β}}_{2,i}{x}_i/k_{2,i}}{\underset{i=1}{\overset{I}{\Σ}}{x}_i{\mathrm{\α}}_i}---\left(4\right)$

In formula,For metal recovery rate；α_iGrade for raw ore i；Denominator represents the amount of metal in raw ore, and molecule represents Amount of metal contained in concentrate；

5. minimize unit cost:

$MinC\left(\stackrel{\→}{X}\right)=Min\frac{\underset{i=1}{\overset{I}{\Σ}}{r}_i{x}_i+C_{other}}{\underset{i=1}{\overset{I}{\Σ}}(1-u_i)x_i/k_{1,i}+\underset{i=1}{\overset{I}{\Σ}}{u}_i{x}_i/k_{2,i}}+C_{energy}---\left(5\right)$

In formula,Unit cost (yuan/ton)；r_iUnit price (yuan/ton) for raw ore i；C_otherRepresent quota expense, C_energyThe energy consumption cost of representation unit；

The constraints of the production target optimization object function of present embodiment includes: production target target range, set Standby ability, energy resources, tailings grade and available raw ore amount；

1) constraint of production target target range, is primarily referred to as the bound scope of five production targets of ore dressing process about Bundle, the most as follows:

$Q_L\≤Q\left(\stackrel{\→}{X}\right)\≤Q_H---\left(6\right)$

${\mathrm{\β}}_L\≤\mathrm{\β}\left(\stackrel{\→}{X}\right)\≤{\mathrm{\β}}_H---\left(7\right)$

$K\left(\stackrel{\→}{X}\right)\≤K_H---\left(8\right)$

$\mathrm{\ϵ}\left(\stackrel{\→}{X}\right)\≥{\mathrm{\ϵ}}_L---\left(9\right)$

$C\left(\stackrel{\→}{X}\right)\≤C_H---\left(10\right)$

In formula, [Q_L,Q_H] it is the target interval of concentrate yield, Q_LFor lower limit, Q_HFor the upper limit；[β_L,β_H] it is concentrate grade Target interval, β_LFor lower limit, β_HFor the upper limit；K_HThe upper limit is compared for full choosing；ε_LFor metal recovery rate lower limit；C_HFor the unit cost upper limit；

Constraint (6) ensure that concentrate yield, in bound range constraint, retrains (7) and ensure that concentrate grade scope, about Bundle (8) ensure that full choosing ratio is less than given upper limit requirement, and constraint (9) is used for avoiding too much metal loss, constraint (10) use Carry out guarantor unit's cost less than predetermined upper limit requirement.

2) constraint of capacity of equipment: dressing Production Process is limited by various apparatus and process abilities, mineral processing production equipment is put down All during platform, treating capacity also has the constraint of the upper limit.There are five kinds of main main equipments in ore dressing plant, is shaft furnace (j=1), high intensity magnetic separation respectively During ball mill (j=2), the ball mill (j=3) during low intensity magnetic separation, intensity magnetic separator (j=4), filter (j=5), The constraint of its capacity of equipment is as follows:

$q_j\left(\stackrel{\→}{X}\right)=\frac{\underset{i=1}{\overset{I}{\Σ}}{u}_i{x}_i}{\underset{k=1}{\overset{K}{\Σ}}{N}_{k,j}{T}_k}\≤q_{j,H},k\∈\{1,2,...K\},j=1---\left(11\right)$

$q_j\left(\stackrel{\→}{X}\right)=\frac{\underset{i=1}{\overset{I}{\Σ}}(1-u_i)x_i}{\underset{k=1}{\overset{K}{\Σ}}{N}_{k,j}{T}_k}\≤q_{j,H},k\∈\{1,2,...K\},j=2---\left(12\right)$

$q_j\left(\stackrel{\→}{X}\right)=\frac{{\mathrm{\η}}_b\underset{i=1}{\overset{I}{\Σ}}{u}_i{x}_i}{\underset{k=1}{\overset{K}{\Σ}}{N}_{k,j}{T}_k}\≤q_{j,H},k\∈\{1,2,...K\},j=3---\left(13\right)$

$q_j\left(\stackrel{\→}{X}\right)=\frac{\underset{i=1}{\overset{I}{\Σ}}(1-u_i)x_i}{\underset{k=1}{\overset{K}{\Σ}}{N}_{k,j}{T}_k}\≤q_{j,H},k\∈\{1,2,...K\},j=4---\left(14\right)$

$q_j\left(\stackrel{\→}{X}\right)=\frac{Q\left(\stackrel{\→}{X}\right)}{\underset{k=1}{\overset{K}{\Σ}}{N}_{k,j}{T}_k}\≤q_{j,H},k\∈\{1,2,...K\},j=5---\left(15\right)$

In formula (11)-(15), N_k,jWhen production series number is k, the operation number of units of jth kind equipment；T_kIn week certain time The production run time of (e.g., the moon or year) series number k in phase；q_j,HThe disposal ability upper limit (ton/little during the average platform of jth kind equipment Time)；N_k,jT_kRepresent the total operation time of jth kind equipment when series number is k that produces.In formula (11), ∑ u_ix_iRepresent time Between ∑ N_k,jT_kTotal lump ore amount that interior shaft furnace processes；In formula (12), ∑ (1-u_i)x_iRepresent at time ∑ N_k,jT_kInterior high intensity magnetic separation mistake The treating capacity of the fine ore that strong magnetic ball mill in journey is total；In formula (13), η_b∑u_ix_iRepresent at time ∑ N_k,jT_kInterior low intensity magnetic separation During the mine-supplying quantity of total roasted ore of weak magnetic ball mill；In formula (14), ∑ (1-u_i)x_iRepresent at time ∑ N_k,jT_k Total mine-supplying quantity of the intensity magnetic separator during interior high intensity magnetic separation；In formula (15),Represent at time ∑ N_k,jT_kInterior concentration Total concentrate treating capacity of filter in dehydration.

Present embodiment capacity of equipment constraint following formula Unify legislation:

$q_j\left(\stackrel{\→}{X}\right)\≤q_{j,H},k\∈\{1,2,...K\},j\∈\{1,...J\}---\left(16\right)$

(3) constraint of energy resources: energy resources, including water, electricity, gas etc., by can supply be retrained.

$Q_{E,p}\left(\stackrel{\→}{X}\right)=q_{E,p}Q\left(\stackrel{\→}{X}\right)\≤Q_{EH,p},p\∈\{1,2,...,P\}---\left(17\right)$

In formula, q_E,pSpecific consumption for energy p；Q_EH,pThe Maximum Supply Quantity of energy p；4) constraint of tailings grade:

$\begin{array}{c}{\mathrm{\θ}}_l\left(\stackrel{\→}{X}\right)=\[\underset{i=1}{\overset{I}{\Σ}}(1-u_i){\mathrm{\α}}_{1,i}{x}_i-\underset{i=1}{\overset{I}{\Σ}}\frac{(1-u_i){\mathrm{\β}}_{1,i}}{k_{1,i}}{x}_i+(\underset{i=1}{\overset{I}{\Σ}}{u}_i{\mathrm{\α}}_{2,i}{x}_i-{\mathrm{\β}}_w{\mathrm{\η}}_w\underset{i=1}{\overset{I}{\Σ}}{u}_i{x}_i)\\ -\underset{i=1}{\overset{I}{\Σ}}\frac{{\mathrm{\β}}_{2,i}{u}_i{x}_i}{k_{2,i}}\]/\[\underset{i=1}{\overset{I}{\Σ}}(1-u_i)x_i-\underset{i=1}{\overset{I}{\Σ}}\frac{(1-u_i)}{k_{1,i}}{x}_i+{\mathrm{\η}}_b\underset{i=1}{\overset{I}{\Σ}}{u}_i{x}_i-\underset{i=1}{\overset{I}{\Σ}}\frac{u_i}{k_{2,i}}{x}_i\]\\ \≤{\mathrm{\θ}}_{l,H},l=0\end{array}---\left(18\right)$

In formula, α_1,iThe grade of the fine ore obtained after separating for the screened process of raw ore i；α_2,iFor the screened process of raw ore i The grade of the lump ore obtained after separation；η_wFor barren rock rate, roasting process the barren rock amount the produced ratio to lump ore amount；β_wIt is useless Stone grade；η_bFor roasting productivity, roasting process the roasted ore amount the produced ratio to lump ore amount；θ_l,HRepresent on tailings grade Limit；

Avoiding too much metal loss by arranging the tailings grade upper limit, we use using lower inequality as mine tailing product The Unified Expression of position constraint.

${\mathrm{\θ}}_l\left(\stackrel{\→}{X}\right)\≤{\mathrm{\θ}}_{l,H}---\left(19\right)$

5) constraint of available ore resource: the constraint that various ore resource amounts usable have bound is as follows:

Q_i,min≤x_i≤Q_i,max, i=1,2 ... I (20)

In formula, [Q_i,min,Q_i,max] be in the plan phase raw ore i can supply interval, Q_i,minFor raw ore i in the plan phase Can the upper limit of supply, Q_i,maxFor in the plan phase raw ore i can the lower limit of supply；

Described production target optimum results output unit includes: optimum results output module, production target monitor mould in real time Block, raw ore sort configuration amount and issue module；

Described optimum results output module is used for output in real time and display by obtaining after the optimization of many production targets prioritization scheme To raw ore sort configuration amount optimal value and production target optimal value；Described raw ore sorts configuration amount and refers to various raw ore money Source, such as non-preliminary election speculum iron, preliminary election speculum iron, Hei Goukuang, local ore deposit, blackhawk mountain lump ore and the consumption in blackhawk SHANFEN ore deposit；

Described production target real-time monitoring module is for showing the real-time of production target actual value and production target optimal value Variation tendency, it is simple to operator check the historical data of dressing Production Process, as the reference adjusting the production schedule；

Described raw ore sorts configuration amount and issues module for by optimum results output module and lower floor's mineral processing production system OPC (the OLE for Process Control) communication mechanism set up, sorts configuration amount optimal value by corresponding raw ore and is issued to Lower floor's mineral processing production system；

Described production target optimizes System Management Unit, including: subscriber information management module, Password Management module, log in/ Exit management module and help module；Described subscriber information management module is used for managing user information, including: rights management, use Family management and log management；Described rights management includes that authority, role definition reclaim with role；Described user management includes user Creating, change, delete, authority, role give and reclaim；Carried out after described log management, i.e. record user's entrance system Important data should be had recovery function by operation；

Described Password Management module is used for managing user cipher；Native system uses password to take off control to manage system, the closeest Code is managed by user oneself, and other user includes that system manager must not change user cipher, but system manager can delete Except this user；

Described log in/exit module: be used for managing user and log in/exit process；Complete user and log in associative operation, such as power Limit certification etc.；

Described help module is for according to user's needs, and the system that provides the user uses help information.

The raw ore based on man-machine interaction of present embodiment sorts process many production targets optimization method, the base described in employing Raw ore in man-machine interaction sorts the many production targets of process and optimizes system realization, as in figure 2 it is shown, comprise the following steps:

Step 1: determine many production targets Optimized model, according to many production targets Optimized model, optimizes in many production targets In algoritic module storehouse, the many production targets optimized algorithm needed for searching according to algorithm parameter configuration, if not finding, then perform step 2, if finding, then perform step 3；

Present embodiment preferred many production targets optimized algorithm is non-dominated sorted genetic algorithm based on gradient information G-NSGA-II。

Step 2: required many production targets optimized algorithm is packaged into new many production targets optimization algorithm module, and handle This new many production targets optimization algorithm module is added in many production targets optimization algorithm module storehouse, and performs step 1；

Step 2.1: by storing path, sets up many production targets optimized algorithm to be packaged and multi-objective optimization algorithm mould Corresponding relation between block storehouse；

Step 2.2: be provided for reading the algorithm parameter configuration interface of different language, it includes the input and output ginseng of algorithm Number, the title of algorithm；

Step 2.3: algorithm is preserved to multiple-objection optimization algoritic module storehouse by the algorithm parameter configuration interface according to arranging；

Step 3: the parameter needed for acquisition；

Respectively in production target desired value module, boundary constraint module, the man-machine interaction of optimized algorithm inner parameter module On interface, on-line setup or revise corresponding parameter, or, from data base, obtain corresponding parameter；

Each production target in ore dressing plant should be the most restricted, as concentrate grade, concentrate yield all have bound to retrain, and entirely selects Then having the upper limit to retrain than with unit Concentrate cost, metal recovery rate has lower limit to retrain.Select from production target desired value Component Gallery Select production target desired value module, by each production target target of corresponding human-computer interaction interface on-line setup present embodiment Value scope, as shown in table 1.

Table 1 production target desired value

Production target title	Desired value
		Comprehensive concentrate grade (%)	[52.5,53.0]
Entirely select than (again)	2.0
		Metal recovery rate (%)	75.0
Unit Concentrate cost (yuan/ton)	200.0
		Comprehensive concentrate yield (ten thousand tons)	[22.8,23.0]

The production facility information in ore dressing plant in present embodiment, including producing number of lines, the key equipment of every production line Having shaft furnace, strong magnetic grinding machine and weak magnetic grinding machine, various equipment have the disposal ability upper limit to retrain, the platform of every production line various kinds of equipment Number restriction and the operation time restriction of each bar production line in this month.

When initially setting, the also quality constraint in reply semi-finished product ore deposit is set, including: the constraint of the tailings grade upper limit, Roasting rate restriction, abandoned mine rate and abandoned mine grade, concentrate consumption；Other constraintss include that the energy constraint consumed is solid with other Determine expense to limit, in the present embodiment, the edge-restraint condition of setting, as shown in table 2.

Table 2 edge-restraint condition

Boundary constraint title	Binding occurrence	Unit
			Shaft furnace ability (q1,H)	25.0	Ton hour
Ball mill ability (q2,H)	80.0	Ton hour
			Ball mill ability (q3,H)	80.0	Ton hour
Sintering consumes concentrate amount (QS)	22.8	Ten thousand tons
			The total tailings grade upper limit (θ0,H)	21.0	%
Roasting rate (ηb)	82	%
			Barren rock rate (ηw)	14	%
Barren rock grade (βw)	14	%
			6 series, the operation number of units (N of the 1st kind equipment6,1)	12	-
8 series, the operation number of units (N of the 1st kind equipment8,1)	15	-
			6 series, the operation number of units (N of the 2nd kind equipment6,2)	3	-
8 series, the operation number of units (N of the 2nd kind equipment8,2)	4	-
			6 series, the operation number of units (N of the 3rd kind equipment6,3)	3	-
8 series, the operation number of units (N of the 3rd kind equipment8,3)	4	-
			6 series operation time (T6)	2.08	My god
8 series operation time (T8)	27.08	My god
			Ton essence energy consumption cost (Cenergy)	17.66	Yuan/ton
Other total Fixed Costs (Cother)	3410000	Unit

Step 4: build many production targets prioritization scheme；

Call inside production target data acquisition module, production target desired value module, boundary constraint module, optimized algorithm Parameter module, optimum results output module, raw ore sort configuration amount and issue module, production target real-time monitoring module and production and refer to Mark optimization algorithm module, builds many production targets prioritization scheme；The prioritization scheme that present embodiment builds is as it is shown on figure 3, produce The outfan of the outfan of target goals value module, the outfan of boundary constraint module and optimized algorithm inner parameter module all connects Delivering a child and produce the input of index optimization algoritic module, the outfan of many production targets optimization algorithm module connects optimum results output The input of module, an outfan of optimum results output module connects production target and issues the input of module, optimizes knot Really another outfan of output module connects an input of production target real-time monitoring module, and production target issues module Outfan connects the input of lower floor's production system, and the outfan of lower floor's production system connects production target data acquisition module, The outfan of production target data acquisition module connects another input of production target real-time monitoring module, operator's reference The data that production target real-time monitoring module obtains, adjust parameter, persistently carry out the optimization of many production targets.

Step 5: many production targets prioritization scheme that operating procedure 4 builds, then obtain at optimum results output module simultaneously Raw ore sorts configuration amount optimal value and production target optimal value as shown in Table 3 and Table 4, obtain in production target real-time monitoring module Single or combination production target actual value and the real-time change tendency information of production target optimal value；

Table 3 production target optimal value

Table 4 raw ore sorts configuration amount optimal value

Actual value and the real-time change trend of optimal value of each production target is shown with the form of trendgram；As shown in Figure 4 Full choosing compare trendgram, it is shown that be i.e. the real-time change trend of actual value and the optimal value entirely selecting ratio；Operator can be led to Cross trendgram and check the historical data of dressing Production Process, as with reference to adjusting the production schedule in time.

Step 6: from the production target optimal value that optimum results output module shows, according to the real-time change of production target value Change tendency information, decisionmaker's preference information and be actually needed, selecting one group of production target optimal value, incite somebody to action as shown in table 5, meanwhile The raw ore that this group production target optimal value is corresponding sorts configuration amount and is issued to lower floor's mineral processing production system；As shown in Table 5, production refers to Mark optimal value in production target target range, inclined between production target actual value and the production target optimal value in ore dressing plant Difference is worth little than the deviation between production target actual value and production target desired value.Therefore, produce in deviation allowed band and refer to Mark optimal value can be as the setting value of ore dressing plant production target.Table 3 lists the production target utilizing prioritization scheme to solve Optimal value, show in the output module is exactly one group of solution therein.

Table 5 comprehensive production index optimum results

During ore dressing optimizes, according to the restriction range of production target, how sorting raw ore configuration amount is that can optimization The key factor being successfully completed, according to the preference information of operator, the optimal value of production target in selection system, therewith Demonstrate raw ore configuration amount target range and raw ore configuration amount optimal value, use with decision-making in actual ore dressing process simultaneously Raw ore configuration amount actual value compares, and facilitates operator to make the reasonable distribution of raw ore configuration amount, and raw ore configuration amount calculates Result is as shown in table 6, for the raw ore configuration amount corresponding with one group of production target optimal value that table 5 selects.

Table 6 raw ore configuration amount result

Raw ore title	Desired value	Optimal value	Actual value	Deviation
					Non-preliminary election essence iron mine (ton)	[60000,80000]	76145	75018	-1127
Preliminary election essence iron mine (ton)	[180000,210000]	199060	208659	9599
					Hei Gou ore deposit (ton)	[80000,100000]	86272	85000	-1272
Local ore deposit (ton)	[50000,70000]	50590	60050	9460
					Blackhawk mountain lump ore (ton)	[5000,10000]	7180	10458	3278
Blackhawk SHANFEN ore deposit (ton)	[30000,40000]	30812	35000	4188

As shown in Table 6, raw ore configuration amount optimal value is in raw ore configuration amount target range, wherein, and departure=reality Value-optimal value.The departure between deviation ratio actual value and desired value between actual value and optimal value is little, therefore raw ore configuration Amount optimal value, can be as the setting value of on-the-spot raw ore configuration amount in deviation allowed band.

Step 7: judge that whether production target actual value that production target data acquisition module shows is in production target target In the range of value, it is then to perform step 6, no, then perform step 3.

Present embodiment preferred production target optimized algorithm is non-dominated sorted genetic algorithm G-based on gradient information NSGA-II, quick non-dominated sorted genetic algorithm inner parameter based on gradient information arranges and includes: population number, iteration time Number, object function number, decision variable number, constraints number, crossover probability and mutation probability, cross and variation distribution is Number, gradient-driven probability, as shown in table 7.

Table 7 quick non-dominated sorted genetic algorithm inner parameter based on gradient information is arranged

This algorithm is as it is shown in figure 5, specifically comprise the following steps that

Step A: randomly generate initial parent population P₀, then to P₀In all individualities carry out non-dominated ranking, and respectively A fitness value is distributed for each individuality, then to P₀Perform binary algorithm of tournament selection, intersect, make a variation, obtain new filial generation kind Group Q₀；

Step B: form new colony R_t=P_tUQ_t, wherein iterations t=0,1,2 ..., gen；Population R_tCarry out non- Join sequence, obtain non-dominant forward position F₁, F₂..., F_i, i=1,2 ...；

Step C: to all non-dominant forward position F_iCrowding compares operation sequencing, and selects the most best N number of Body forms population P_(t+1)；

In non-dominated sorted genetic algorithm based on gradient information, Sharing Function need to be specified to guarantee population by policymaker Multiformity, in order to solve this problem, it is proposed that crowding concept, crowding is individual around the set point in population Density.

After non-dominated ranking, need each individual distribution crowding.Owing to each individuality is by level and to gather around Crowded degree carries out selecting, so needing each individual distribution crowding.The distribution of crowding for same level, Therefore comparing crowding between different layers to have little significance, the computation rule of crowding is as follows:

1) crowding for a certain layer boundary point calculates as shown in formula (1):

I[1]_distance=I [l]_distance=∞ (1)

In formula, I is the non-dominant collection in population, and l is the number of population, I [1]_distanceIt is first individual crowding, I[l]_distanceIt is the l individual crowding.

2) crowding of other point calculates as shown in formula (2):

$I{\[i\]}_{dis\tan ce}=I{\[i\]}_{dis\tan ce}+(I\[i+1\].m-I\[i-1\].m)/(f_m^{max}-f_m^{min})---\left(2\right)$

In formula, for each layer of F_i, wherein I [i] .m represents the m-th target function value that in I, i-th is individual,It is The maximum of m object function,Minima for m-th object function.Crowding calculate basic thought be find same In Ceng in m-dimensional space distance between each individuality.

Step D: to population P_(t+1)Perform binary algorithm of tournament selection, intersect and make a variation, and gradient operator, find optimal solution Form population Q_(t+1)；

After all individualities carry out non-dominated ranking and crowding calculates, individuality is selected to enter copulation pond.Copulation pond is selected Selection method uses binary system algorithm of tournament selection, and it is carried out based on tournament selection method.

During binary system algorithm of tournament selection, first randomly choose n individual (wherein n represents the scale of championship), In these individualities, only select one enter copulation pond.Select based on following two criterions:

(1) individuality that Pareto forward position level is smaller is selected；

(2) if two individualities are in the same layer in Pareto forward position, then compare crowding between the two, select crowded Spend bigger individuality.

Non-dominated sorted genetic algorithm uses to intersect and obtains another solution with mutation operator from a solution, and uses matrix Or Pareto dominance relation selects new individuality.Therefore, the population in whole space is random establishment, so produces Population reduces by a target function value to I haven't seen you for ages, and thus obtains individual population of future generation.Owing to which produces a lot of solution, Therefore all of target function value can be increased by intersection and mutation process, it determines the time also can increase therewith.The opposing party Face, the direction of search based on gradient, only in negative half space, can be obtained another from a solution and solve, and the individuality created is forever Do not appear in positive half space, it may therefore be assured that fast convergence.But, the direction of search based on gradient purely is more Local optimum easily occurs, and substantial amounts of Performance Evaluation needs gradient information, so can increase the calculating time.Therefore, based on ladder The mixing of degree intersects, mutation operator can preferably produce new population or creates " filial generation ", and brief description is based on gradient Obtain the process of new filial generation:

One n-ary function f (x₁,x₂,…,x_n) can be considered vector variable (x₁,x₂,…,x_n)^TReal-valued function, be denoted as f (x).Definition sets f:Rⁿ→ R (n-dimensional space is mapped as the one-dimensional space),(Belong to the variable in n-dimensional space), if f (x) PointPlace is for independent variable x ∈ (x₁,x₂,…,x_n)^TThe partial derivative of each componentI=1,2 ..., n, all exist, then Claim function f at pointPlace's single order can be led, and vectorFor f (x) at pointSingle order Derivative or gradient.

If function is multiple objective function, then definition sets f:Rⁿ→ R,Note f (x)=(f₁(x),f₂(x),…,f_n (x))^T

If f_i(x) (i=1,2 ..., m) at pointPlace is for independent variable x=(x₁,x₂,…,x_n)^TThe local derviation of each component NumberAll exist, then claim vector function f at pointPlace is that single order can be led, and claims matrix

${\▿}_{m\×n}f\left(\stackrel{\‾}{x}\right)=\left[\begin{array}{c}{(\▿f_1)}^T\\ {(\▿f_2)}^T\\ .\\ .\\ .\\ {(\▿f_m)}^T\end{array}\right]=\left[\begin{array}{cccc}\frac{\∂f_1\left(x\right)}{\∂x_1}& \frac{\∂f_1\left(x\right)}{\∂x_2}& \mathrm{...}& \frac{\∂f_1\left(x\right)}{\∂x_n}\\ \frac{\∂f_2\left(x\right)}{\∂x_1}& \frac{\∂f_2\left(x\right)}{\∂x_2}& \mathrm{...}& \frac{\∂f_2\left(x\right)}{\∂x_n}\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ \frac{\∂f_m\left(x\right)}{\∂x_1}& \frac{\∂f_m\left(x\right)}{\∂x_2}& \mathrm{...}& \frac{\∂f_m\left(x\right)}{\∂x_n}\end{array}\right]$

For f (x) at pointThe first derivative at place, or claim Jacobi matrix, it is abbreviated as

As a example by function of a single variable, for minf (x), wherein f:Rⁿ→ R has the continuous local derviation of single order.Assume iteration K time, kth time iteration point is x_k, andTake the direction of searchFor making target function value at an x_kPlace obtains Obtain the fastest decline, can be along d_kCarry out linear search.

The direction of search can be constructed, such as following formula by above definition:

$d=-\underset{i=1}{\overset{m}{\Σ}}{\mathrm{\λ}}_i\frac{\▿f_i\left(x\right)}{\left|\right|\▿f_i\left(x\right)\left|\right|}$

$e=\frac{d}{\left|\right|d\left|\right|}=-\frac{\underset{i=1}{\overset{m}{\Σ}}{\mathrm{\λ}}_i\frac{\▿f_i\left(x\right)}{\left|\right|\▿f_i\left(x\right)\left|\right|}}{\left|\right|\underset{i=1}{\overset{m}{\Σ}}{\mathrm{\λ}}_i\frac{\▿f_i\left(x\right)}{\left|\right|\▿f_i\left(x\right)\left|\right|}\left|\right|}$

Wherein d is the direction of search, and e is the unit direction of search, λ_i∈ (0,1), works as λ_iDuring ＞ 0, for negative gradient direction.

After determining the direction of search, should determine that step-length τ of search, take step-length τ_kSo that can make in the direction of search of e Quickly find optimal solution.In order to reduce amount of calculation, define τ_k=τ₀(gen-k+1)/gen is optimal step size.Wherein τ_kFor kth The step-length that generation evolves, τ₀Initial step length during for evolving, the situation being typically based on reality gives initial step length, and gen is maximum Evolutionary generation.

$f(x_k+{\mathrm{\τ}}_k{d}_k)=\underset{\mathrm{\τ}\≥0}{min}f(x_k+{\mathrm{\τd}}_k)$

Obtain+1 iteration point x of kth_k+1=x_k+τ_kd_k.Then point range x is obtained₀,x₁,x₂..., wherein x₀For initial point, as ReallyThen x_kBeing the stationary point of f, this is to terminate iteration.

Step E: judge that iterations, whether equal to gen, is then to terminate, no, then t=t+1, go to step B.

Although the foregoing describing the detailed description of the invention of the present invention, but the those skilled in the art in this area should managing Solving, these are merely illustrative of, and these embodiments can be made various changes or modifications, without departing from the principle of the present invention And essence.The scope of the present invention is only limited by the claims that follow.

Claims (3)

1. a raw ore based on man-machine interaction sorts process many production targets optimization system, it is characterised in that: including: produce and refer to Mark data acquisition module, parameter set and amendment unit, many production targets optimization algorithm module storehouse, many production targets Optimized model Module, many production targets optimum results output unit and many production targets optimize System Management Unit；

Described production target data acquisition module is for by the OPC communication mechanism with lower floor's mineral processing production system foundation, gathering Many production targets actual value of lower floor's mineral processing production system also shows；Described many production targets, including: concentrate grade, complete Choosing ratio, metal recovery rate, unit Concentrate cost and concentrate yield；

Described parameter sets and amendment unit, including: inside production target desired value module, boundary constraint module, optimized algorithm Parameter module；These modules all carry out man-machine interaction by corresponding human-computer interaction interface；Described parameter, including: produce refers to more Heading scale value, edge-restraint condition and production target optimized algorithm inner parameter；

Described production target desired value module, is used for the target range to each production target on corresponding human-computer interaction interface It is set and revises；The target range of described production target, including: concentrate grade bound, entirely select the upper limit of ratio, gold Belong to the lower limit of the response rate, the upper limit of unit Concentrate cost and the bound of concentrate yield；

Described boundary constraint module, for sorting the boundary constraint bar that process relates on corresponding human-computer interaction interface to raw ore Part is set and revises；Described boundary constraint, including: production target target range, capacity of equipment, tailings grade, available Energy resources, available ore resource；Wherein production target target range is not set in this module and revises, but logical Cross and call the acquisition of production target desired value module；

Described optimized algorithm inner parameter module, for excellent to the production target of user's design on corresponding human-computer interaction interface The inner parameter changing algorithm is configured and revises；Described inner parameter, the parameter needed for algorithm itself；

Described many production targets optimization algorithm module storehouse, the collection constituted for multiple different many production targets optimization algorithm module Close；

Described many production targets optimization algorithm module, the many production targets optimized algorithm determined for encapsulation, will each give birth to more Producing index optimization algorithm packaging is production target optimization algorithm module more than；

Described many production targets Optimized model module, is used for loading many production targets Optimized model；

Described many production targets optimum results output unit includes: optimum results output module, production target real-time monitoring module, Raw ore sorts configuration amount and issues module；

Described optimum results output module in real time output and display by obtaining after the prioritization scheme optimization of many production targets Raw ore sorts configuration amount optimal value and production target optimal value；Described raw ore sorts configuration amount and refers to disappearing of various ore resource Consumption；

Described production target real-time monitoring module is for showing the real-time change of production target actual value and production target optimal value Trend；

Described raw ore sorts configuration amount and issues module for by optimum results output module and lower floor's mineral processing production system foundation OPC communication mechanism, corresponding raw ore is sorted configuration amount optimal value and is issued to lower floor's mineral processing production system；

Described production target optimizes System Management Unit, including: subscriber information management module, Password Management module, log in/exit Management module and help module；Described subscriber information management module is used for managing user information, including: rights management, Yong Huguan Reason and log management；Described Password Management module is used for managing user cipher；

Described log in/exit module: be used for managing user and log in/exit process；

Described help module is for according to user's needs, and the system that provides the user uses help information.

2. raw ore based on man-machine interaction sorts process many production targets optimization method, described in employing based on man-machine interaction Raw ore sort the many production targets of process optimize system realize, it is characterised in that: comprise the steps:

Step 1: determine many production targets Optimized model, according to many production targets Optimized model, at many production targets optimized algorithm In module library, the many production targets optimized algorithm needed for searching according to algorithm parameter configuration, if not finding, then perform step 2, if Find, then perform step 3；

Step 2: required many production targets optimized algorithm is packaged into new many production targets optimization algorithm module, and this is new Many production targets optimization algorithm module add in many production targets optimization algorithm module storehouse, and perform step 1；

Step 3: the parameter needed for acquisition；

Respectively at production target desired value module, boundary constraint module, the human-computer interaction interface of optimized algorithm inner parameter module On, on-line setup or revise corresponding parameter, or, from data base, obtain corresponding parameter；

Step 4: build many production targets prioritization scheme；

Call production target data acquisition module, production target desired value module, boundary constraint module, optimized algorithm inner parameter Module, optimum results output module, raw ore sort configuration amount and issue module, production target real-time monitoring module, many production targets Optimization algorithm module, builds many production targets prioritization scheme；

Step 5: many production targets prioritization scheme that operating procedure 4 builds, then obtain raw ore at optimum results output module simultaneously Sort configuration amount optimal value and production target optimal value, obtain production target optimal value and life in production target real-time monitoring module Produce the real-time change tendency information of index actual value；

Step 6: from the production target optimal value that optimum results output module shows, become according to the real-time change of production target value Gesture information, decisionmaker's preference information and be actually needed, select one group of production target optimal value, meanwhile, by excellent for this group production target Raw ore corresponding to change value sorts configuration amount optimal value and is issued to lower floor's mineral processing production system；

Step 7: judge that whether production target actual value that production target data acquisition module shows is at production target desired value model In enclosing, it is then to perform step 6, no, then perform step 3.

Raw ore based on man-machine interaction the most according to claim 2 sorts process many production targets optimization method, its feature It is, required many production targets optimized algorithm is packaged into new many production targets optimization algorithm module by described step 2, The method of described encapsulation, comprises the steps:

Step 2.1: by storing path, sets up many production targets optimized algorithm to be packaged and many production targets optimized algorithm mould Corresponding relation between block storehouse；

Step 2.2: be provided for reading the algorithm parameter configuration interface of different language, it includes the input/output argument of algorithm, The title of algorithm；

Step 2.3: algorithm preserved at most in production target optimization algorithm module storehouse according to the algorithm parameter configuration interface arranged.