JP4612389B2 - Production and / or distribution plan creation device and method, process control device and method, and computer program - Google Patents

Description

The present invention is suitable for production and / or logistics (in the present specification, production) using, for example, a transport process in which a plurality of products are transported through a plurality of different transport paths and each product can select a plurality of different transport paths. (Also described as physical distribution) The present invention relates to a plan creation device and method, a process control device and method, and a computer program.

  Conventionally, in manufacturing processes in many industries including steel, a plurality of products are processed in a plurality of different processes, and each product is manufactured under a condition that a plurality of different process paths can be selected. For example, in the ironmaking raw material transfer process from the yard to the storage tank at a steelworks, multiple brands were stacked for multiple tanks with different brands entered and discharged at different cutting speeds. Select a pile that matches the brand in the tank from multiple yards, select an appropriate reclaimer from among the available reclaimers, perform the cutting process from the pile with the reclaimer, An appropriate belt conveyor system is selected from a plurality of belt conveyor systems that can be transported, transported to a storage tank, and an appropriate tank amount is stored in the storage tank from an appropriate start time to an end time. In this case, a plurality of brands are stacked in a plurality of yards, and the same brand may be stacked in a plurality of mountains.

  When planning the production of steelmaking raw materials from yard to storage tank, take into account the operational restrictions of the yard, reclaimer, belt conveyor series, storage tank, and restrictions caused by the raw material logistics process. It is necessary to put in. That is, in the raw material factory, the outage of the storage tank should be kept to a minimum or should not occur in order to stabilize the operation of the blast furnace operation and the sintering factory. To this end, it is necessary to constantly monitor the inventory changes of many mining tanks and to be always careful.

  In addition, when the stock level of the storage tank falls below a certain level, the coarse iron ore that has accumulated in the tank flows out at once, causing the iron ore to become unstable in grain size and falling when entering the sintered ore. There is a possibility that pulverization may occur because the distance becomes large. In order to prevent these, the storage tank inventory is required to be highly stable.

  Furthermore, not only the process route is different for each brand entering the storage tank, but there are multiple process paths even when entering the same storage tank, so the use status of the equipment is judged and an appropriate process route is selected. There is a need to.

  Furthermore, it is necessary to take into consideration that the processing time in each process and each equipment is different.

  Under such various restrictions, if one reclaimer is used from one mountain, it is transported by one belt conveyor system, and it enters into one storage tank, the storage tank is simply If the inventory level is low, iron ore can be transported from the mountain to the storage tank at that time.

  However, as described above, multiple brands are stacked in multiple yards, and multiple brands are stacked in different mountains. In order to improve the quality, it is necessary to make a raw material yard operation plan for which storage tanks should be used in what order and which reclaimer and belt conveyor system will be used, and from when to when the tanks will be installed. In order to stabilize the operation of the sinter plant, it is possible to secure inventory, stabilize the iron ore particle size, and prevent the sinter ore from being pulverized.

  In general, in a manufacturing process in which a plurality of products are processed and manufactured in a plurality of different processes, production / distribution planning in the process or automation of distribution control is desired. Conventionally, as a technique for this automation, a discrete event simulator or the like by various methods has been proposed. For example, as disclosed in “Production Plan Evaluation Method and System” of Patent Document 1, on a simulator simulating a factory built on a computer, information obtained from the same interface as the actual device is used. By predicting the operation of the product, performing virtual production at a faster speed than the actual equipment based on the operation prediction, and presenting a highly accurate index using the virtual production process and results, the production plan It is a technique that enables evaluation and selection.

JP 2002-366219 A

  As disclosed in “Production Plan Evaluation Method and System” described in Patent Document 1 above, the method of creating a production / distribution schedule using a conventional discrete event simulator requires a satisfactory result. (1) It was necessary to perform simulation while changing the conditions in various ways, and repeatedly evaluate the results. Therefore, (2) a large-scale factory has a problem that it takes a lot of time to create a production / distribution schedule. In addition, (3) in order to obtain a highly accurate production / distribution schedule, there is a problem that the simulation rules must be set in detail.

  Furthermore, in the conventional method, a process is possible in which a plurality of products are processed in a plurality of different processes, and each product can select a different multi-process path (even if different equipment is used in the same process). When considered, considering all selectable routes, the number of cases was too large to be calculated. For this reason, as in the above-mentioned document, various conditions are changed, and the route that is considered to be good is selected from among the selected routes. In addition, if optimization is performed by a conventional method, there is a problem that analysis time becomes long.

Accordingly, the present invention considers a plurality of production and / or logistics objects to be processed in different multi-process paths, and each object has a different multi-process path (even when different equipment is used in the same process, as different process paths). ) Can be selected at high speed, and production and / or distribution planning or distribution control matching the operation plan can be optimized at high speed under a given operation constraint.

The production and / or distribution plan creation apparatus, method, and computer program according to the present invention process a plurality of production and / or distribution objects in different multi-process paths, and select different multi-process paths for each object. in intended to create a production and / or logistics planning in can production and / or distribution process, the production and / or actual operation on production logistics processes and / or logistics flow and production and / or logistics constraints model And a simulator that detects production and / or distribution status and production and / or distribution restrictions when an event occurs, and a target period (plan creation period) set in advance from the production start date and time of the production and / or distribution plan. ) content as a target of, among the aforementioned production and / or distribution process of production and / or distribution conditions and production and / or logistics constraints, or rough A mathematical model holding device that holds a mathematical model consisting only of the set part, an evaluation function setting device that sets an evaluation function for evaluating the mathematical model held by the mathematical model holding device, and the mathematical model holding device And the optimization calculation device that performs an optimization calculation process using the mathematical model held by the evaluation function and the evaluation function set by the evaluation function setting device, and calculates a physical distribution instruction for the simulator. By the optimization calculation processing by the computing device, the distribution instruction is calculated for the target period (instruction calculation period) set within a predetermined range from the planning start date and time, and is given to the simulator, and the target period selected in advance The simulation is executed for (simulation period), and the above simulation is performed for a preset period (plan finalization period). To confirm the Shon resulting production and / or logistics planning, simulation obtained by repeating the process of devising the production and / or logistics planning and setting the date and time immediately after a new design starting date of the definite time period A production and / or logistics plan in the production and / or logistics process is created from the results, and the mathematical model is constructed for all patterns or partial patterns of the combination process that can be selected as the process route, and It is characterized in that an optimization calculation process by the optimization calculation apparatus is performed on each constructed mathematical model, and an optimal one is selected.
Further, the process control apparatus, method, and computer program according to the present invention process a plurality of production and / or physical distribution objects in different multi-step paths, and can select a multi-step path in which each target object is different. In order to control the logistics process, the production and / or logistics process and / or logistics flow and production and / or logistics constraints of the above production and / or logistics process are modeled and production when an event occurs. and / or the distribution condition and simulator for detecting the production and / or logistics constraints, targeting the planning starts preset period from time (planning period) content of the production and / or logistics planning, the production and / or logistics process production and / or logistics state of the production and / or logistics constraints, a mathematical model consisting of only a part which is set in advance A mathematical model holding device to hold, an evaluation function setting device for setting an evaluation function for evaluating the mathematical model held by the mathematical model holding device, a mathematical model held by the mathematical model holding device, and the evaluation The optimization calculation processing is performed using the evaluation function set by the function setting device, and the optimization calculation device is used to calculate the distribution instruction for the simulator. The distribution instruction is calculated for the target period (instruction calculation period) set within a predetermined range from the start date and time, and is given to the simulator, and the simulation is executed for the target period (simulation period) selected in advance. , producing the simulation results for a preset time period (schedule determined period) content and / or distribution plan And to confirm, the finalized based on when the set as a new design starting date production and / or logistics planning simulation obtained by repeating the process of devising the result immediately after the period the production and / Alternatively, the optimization calculation device is configured to control the physical distribution process, construct the mathematical model for all patterns or a part of patterns of the combination processes that can be selected as the process route, and perform the optimization calculation device for each of the constructed mathematical models. It is characterized in that the optimum calculation is selected by performing the optimization calculation process according to.

According to the present invention, a plurality of production and / or physical distribution objects are processed in different multi-process paths, and each object has a different multi-process path (even when different equipment is used in the same process, In the process that can select (Thinking), the formula model is constructed for all or a part of the combination process that can be selected as the process path, and the optimized computation processing is performed by the optimization computing device for each of the constructed formula models. In order to select the most suitable one, even in the case of complicatedly intertwined processes, even when the best selection of the process to process the product is necessary, the operation plan is subject to the given operation constraints. Matched production and / or logistics planning or logistics control can be optimized at high speed.

  Hereinafter, a production / distribution plan creation apparatus and method, a process control apparatus and method, and a computer program according to the present invention will be described with reference to the drawings. First, the basic configuration of the production / distribution plan creation apparatus of this embodiment will be described with reference to FIG.

  FIG. 10 is a block diagram illustrating an example of the overall configuration of a production / distribution plan creation apparatus to which the present invention is applied. As shown in FIG. 10, the production / distribution plan creation apparatus 300 of this embodiment includes a discrete event system simulator 310, a mathematical model holding apparatus 320, an optimization calculation apparatus 330, an evaluation function setting apparatus 340, and the like. .

  The discrete event simulator 310 is a large discrete event simulator simulating a factory or the like, and is configured as a discrete event system that moves an object for each event (an event of the discrete event simulator). In the present embodiment, the discrete event simulator 310 is configured using a Petri net, and a mathematical model 321 is output. Here, the mathematical model is output by the number of all or some of the selectable combination processes.

  Further, a physical distribution model (mathematical model) 321 is configured in correspondence with the discrete event system simulator 310. In this embodiment, product acceptance plan, product shipment plan, inventory plan, equipment use plan, equipment repair plan, equipment capacity, process status, equipment status, inventory status, equipment operation / failure status in manufacturing process / conveyance, and Based on input data representing all or part of the operational prerequisites from the operator, the product is based on the accuracy set in advance for the target period set in advance from the planning start date and time of the production / distribution plan. The mathematical model 321 is constructed with respect to the relations and constraints of the work groups accompanying the processing of the moving body and the equipment.

  The mathematical model 321 is held by a mathematical model holding device 320 configured by semiconductor storage means or the like. Then, optimization calculation is performed by the mathematical model 321 and the optimization calculation device 330, and a physical distribution instruction for the discrete event simulator 310 is calculated. The optimization calculation performed by the optimization calculation device 330 is performed using the evaluation function S.

  Therefore, according to the production / distribution plan creation apparatus of the present embodiment, the distribution calculation is not performed based on a predetermined rule as in the prior art, but the optimum calculation performed by the optimization calculation apparatus 330 is performed. A physical distribution instruction based on the result can be output to the discrete event simulator 310. As a result, it is possible to reliably issue an optimum physical distribution instruction according to the event at that time.

  In addition, when a new event occurs, a calculation instruction is output from the discrete event system simulator 310 so that the mathematical model 321 and the optimization calculation device 330 perform calculations. When the calculation instruction is given from the discrete event simulator 310, the optimization calculation device 330 executes optimization calculation using the mathematical model 321 and the evaluation function S. As described above, an optimal production / distribution schedule can be created by executing a detailed simulation in which the discrete event simulator 310 and the optimization calculation device 330 are linked for each event once.

  That is, the simulation performed in the present embodiment is not based on the simulation based on a predetermined rule as in the prior art, but is performed based on the result of the optimal calculation, so only one simulation is performed. This makes it possible to reliably obtain the theoretical optimal solution, and it is not necessary to evaluate the simulation result and repeat the simulation many times as in the past, and to create the simulation result 350 quickly and with high accuracy. Can do. Therefore, even if the schedule is created on a large scale, it can be sufficiently created within a practical time. The simulation result 350 obtained as described above is output as a schedule.

  In addition, even if the scale of the discrete event simulator 310 is very large, or the constraint conditions are very large and complicated, it affects the schedule creation among the distribution states and mathematical formulas described in the discrete event simulator 310. By taking only the important part having a large value into the mathematical model 321, the scale of the discrete event simulator 310 can be set within an appropriate range, and the optimization calculation can be performed within a practical time. it can.

  Since the discrete event system simulator 310 can describe all the physical distribution states and physical distribution constraints to be considered, it is guaranteed that the schedule created by performing one simulation can be actually executed.

  As described above, in the present embodiment, since the discrete event simulator 310, the mathematical model 321 and the optimization calculation device 330 are linked to create a physical distribution schedule, (1) repetition of simulation You can create a schedule without In addition, (2) it is possible to reduce the calculation time by incorporating only important parts that have a large influence on schedule creation into the mathematical model 321 and (3) it is possible to solve large-scale problems. .

  Further, every time an event requiring a distribution instruction occurs, the distribution state and distribution constraint information of the discrete event simulator 310 is detected, and the optimization calculation is performed based on the detected information and a predetermined evaluation index. The apparatus 330 calculates an optimal logistics instruction by an optimization method and creates a schedule by performing a detailed simulation by the discrete event system simulator 310 based on the calculation result, so that (4) the schedule accuracy can be increased. (5) It is possible to create a schedule whose feasibility is verified.

  In addition, since the mathematical model 321 is introduced, it is possible to quickly cope with a change in an important part that has a large influence on schedule creation, and a schedule creation device with high maintainability can be constructed.

  Then, the mathematical model 321 is created by taking in the information related to creating the schedule of the physical distribution of interest for the period (plan creation period) set in advance from the planning start time of the production / distribution process. Then, the created mathematical model 321 is given to the optimization calculation device 330, and a logistics instruction is calculated for a period (instruction calculation period) preset from the present time by the optimization calculation processing by the optimization calculation device. This is given to the discrete event simulator 310, and the simulation is executed for a preset period (simulation period), and the distribution plan is determined for a preset period (plan determination period).

  Next, the production / distribution schedule in the production / distribution process is obtained from the simulation result 350 obtained by repeating the process of setting the date and time immediately after the determined period as the new planning start date and time and planning the distribution plan. I try to create it.

(First embodiment)
Hereinafter, a first embodiment of the present invention will be described with reference to the drawings. In the production / distribution plan creation device of this embodiment, in the transfer from the yard to the storage tank as described in the conventional example, the yard loading brand, the yard inventory amount transition, the iron ore / sinter cut out amount, the equipment layout In order to stabilize the operation of blast furnaces and sintering plants under the constraints of raw material logistics, etc., the stock level was secured, the iron ore grain size was stabilized, and the stock level was stabilized to prevent sinter ore from being pulverized. The material yard operation plan optimization problem shall be dealt with. However, this is merely an example, and the production / distribution plan creation apparatus of the present embodiment processes a plurality of products in a plurality of different processes, and each product has a different multi-process route (using different equipment in the same process). It can be applied when creating an operation plan in the target process while keeping many restrictions imposed on the upper and lower processes and the target process. It is possible and particularly effective.

  In the operation plan here, first of all, we secure the stock to stabilize the operation of the blast furnace and sintering plant (prevention of out of stock), stabilize the iron ore particle size, and prevent the sinter ore from being pulverized In order to achieve this, the aim is to achieve a high level of inventory stability.

  In addition, because the amount of cut out differs for each storage tank, and the tank entry conditions, for example, the stock level of the corresponding storage tank, etc. are different at the time of starting the tank entry, the amount to be entered depends on the situation. It is necessary to determine that the level of the storage tank is highly stable.

  In addition, as shown in FIG. 2 which is a schematic diagram of the raw material yard manufacturing process (conveyance) which is an object of the present invention, a plurality of reclaimers and belt conveyor series are used for conveyance to the storage tank to be filled. It is necessary to select an appropriate reclaimer and belt conveyor series because the reclaimer has different cutting ability depending on the reclaimer and there are fewer reclaimers than the number of storage tanks. .

  Under these constraints, when securing the stock of all storage tanks and creating a raw material yard operation plan with a stable inventory level, the order of tank entry, tank start and end times, tank volume, reclaimer It is necessary to accurately determine not only the operation start time and the reclaimer operation end time but also the payout hill, yard, used reclaimer, conveyor belt conveyor series, and incoming tank storage tank.

  FIG. 1 is a block diagram showing a schematic configuration of a production / distribution plan creation apparatus according to the present embodiment, FIG. 2 is a schematic diagram of a raw material yard manufacturing process (conveyance), which is an object of implementation, and FIG. It is a figure which shows the positioning of the production and physical distribution plan preparation apparatus by. First, the positioning of the production / distribution plan creation apparatus according to the present embodiment will be described with reference to FIG.

  As shown in FIG. 3, when creating a raw material yard operation plan, first, in the condition setting and taking-in unit 30, the constraint conditions such as the yard layout, the amount of cut out of the storage tank, and the like that are necessary for formulating the plan, Capability conditions and preconditions are set or imported from the process computer (procone).

  The production / distribution plan creation unit 31 of the present embodiment, under the various distribution constraints set by the condition setting and capture unit 30, the raw material yard operation plan to satisfy these distribution constraints, capacity conditions, etc., that is, Obtain the order of tank entry, tank entry start / end time, reclaimer start / end time, payout hill / yard, reclaimer used, conveyor belt conveyor line, and tank storage tank.

  In this production / distribution plan creation unit 31, as will be described in detail below, a discrete event simulator using a Petri net that graphically models the distribution structure (equipment arrangement in the factory and its connection relationship, facility capacity, process route, etc.) With a completely new combination with mathematical programming methods such as LP (Linear Programming), MIP (Mixed Integer Programming), and QP (Secondary Programming), which are often used as solutions for static programming problems, Optimize processing order, processing time, and process route to be processed from yard to storage tank.

  Raw material yard operation plan obtained by the production / distribution plan creation unit 31 (order of tank entry, tank start / end time, start / end time of reclaimer, payout mountain / yard, used reclaimer, conveyor belt conveyor line, tank storage The information on the mine tank) is given to the display unit 32, and is displayed in, for example, a Gantt chart format and a mine inventory transition graph format.

  The various evaluation units 33 evaluate the obtained operation plan from various viewpoints (for example, inventory transition, continuous issue ability of the same brand in the reclaimer, etc.). Correct entry / exit times of tanks, payout mountains, reclaimers used, etc. Then, the production / distribution plan creation unit 31 creates an operation plan again. Further, at this time, it is possible to fix the tank entry time, the payout pile, the use reclaimer designation, etc. only for the treatment designated as necessary.

  Next, processing performed by the production / distribution plan creation unit 31 will be described. The production / distribution plan creation unit 31 sets the yard layout, process route, tank brand, etc., and distribution constraints, and at each simulation time (every occurrence of the reclaimer operation start event or multiple times of event occurrence). (Each time) Estimate the stock level of the storage tank that is the future logistics state, and optimize the predetermined evaluation function set for avoiding the storage tank stock out of stock and stabilizing the storage tank stock level high The order of tank entry, tank start and end time, tank volume, reclaimer operation start time, reclaimer operation end time, as well as payout mountain, yard, used reclaimer, conveyor belt conveyor line, and tank storage tank are determined. At this time, the estimated range of the future physical condition is set to an appropriate value of about 2 hours or 3 hours from the simulation current time.

  FIG. 4 shows a simple example of the scale reduction of the raw material yard manufacturing process (conveyance) used for explaining the outline of the process, in the production / distribution plan creation part 31 described above. A block diagram (FIG. 1) showing a schematic configuration of the apparatus 31 and FIGS. 5 to 8 showing details of operations inside the production / distribution plan creation section 31 when this example is used will be described in detail.

  In the example of FIG. 4, the yard 1 has a mountain where iron ore brands A, B, and C are stacked, and the yard 2 includes a mountain where brand B is stacked. Reclaimer No. 1 can be used to pay out the yard 1 mountain, and reclaimer No. 2 can be used to pay out the yard 2 mountain. When reclaimer No. 1 is used, iron ore is conveyed by belt conveyor series 1, 2, 3, or 5. When reclaimer No. 2 is used, iron ore is conveyed by belt conveyor series 4 or 6. Is transported. The iron ore transported by the belt conveyor series 1 is stored in the storage tank 1, the belt conveyor series 2, 4 in the storage tank 2, the belt conveyor series 3 in the storage tank 3, and the belt conveyor series 5, 6 in the storage tank. Each is transferred to the tank 4. It is necessary to store the brand A in the storage tank 1, the brand B in the storage tank 2, the brand C in the storage tank 3, and the brand B in the storage tank 4. Here, the brand to be paid out from the yard and the brand to be put into the storage tank must be the same brand.

  In FIG. 1, reference numeral 21 denotes a discrete event simulator based on a Petri net model, which includes a graphical physical distribution structure model based on a Petri net and a rule description that cannot be expressed graphically.

  Here, the Petri net model is a discrete event such as logistics by connecting transitions represented by vertical bars and places represented by circles with arcs represented by arrows and moving black circle tokens between places. It is a technique for modeling a system. For details on this Petri Net, please refer to “Theory and Practice of Petri Net (Mikio Aoyama et al., Asakura Publishing)”.

  An example of the example shown in FIG. 4 expressed by a Petri net model is shown in FIG. For example, a place 901 representing the mountain 1 in the yard 1, a place 905 representing the No. 1 reclaimer, a place 907 representing the belt conveyor series 1, a place 913 representing the storage tank 1, and a conveyance judgment from the mountain 1 to the No. 1 reclaimer. Petri as a transition 951 representing transition, transition 955 representing transportation judgment from No. 1 reclaimer to belt conveyor series 1, transition 961 representing transportation judgment from belt conveyor series 1 to storage tank 1, and connection between them by arc Expressed with a net model.

  Here, JOB being processed with No. 1 reclaimer represented as token 971 (considering a series of operations entering the storage tank from the yard in one storage tank processing as one division ) Process route determination rule is set as the firing judgment of the transport judgment transitions 955, 956, 957, and 958 for judging that the passing process list among the token attributes set in the token is transferred from the No. 1 reclaimer to the belt conveyor. To do. In this example, since the belt conveyor system 1 is registered in the passing process list as a subsequent process of the No. 1 reclaimer, the transition 955 can be ignited.

Also, rule description such as time judgment that cannot be described graphically is described using C language or the like such as if and then in the transition. For example, in the token 971, the process start time in each process registered in the current time and process start time list is determined to determine whether firing is possible.
If (current time> process start time in process start time list)
else Cannot fire If the current time is 10:23 of H12 / 3/1, the process processing start time of the belt conveyor series 1 in the processing start time list is 10:21 of H12 / 3/1. It can be ignited.

  The above-described graphical structure model and all conditions of the rule description are judged, and if all of them can be ignited, the token moves to the belt conveyor series 1 which is the next process.

  As described above, the Petri net model of the entire production factory can be constructed by describing the Petri net model in the same manner for all the processes.

  Reference numeral 27 denotes an optimization unit, which is set in the simulator 21 such as the yard layout, process route, tank entry brand setting conditions, logistics constraint information, and the current logistics given as a result of the simulation performed by the simulator 21. The situation (information of current storage tank inventory level, equipment usage status, etc.) is taken in, and optimization calculation is performed for the prediction range (target period (instruction calculation period)) of about 2 hours or 3 hours from the simulation current time Execute, optimal tank order for the above range, tank start and end time, tank volume, reclaimer operation start time, reclaimer operation end time as well as payout mountain, yard, used reclaimer, conveyor belt conveyor series The tank storage tank is determined, and based on this determination, a physical distribution instruction is issued to the simulator 21.

  It is assumed that a simulation is performed by the simulator 21 and a reclaimer operation start event of JOB having the latest reclaimer operation start time among the JOBs determined by the previous optimization calculation at the simulation time t (simulation current time) occurs.

  When such an event occurs, the optimization unit 27 receives this information and first sets a future prediction range as shown in FIG. Extract the storage tank to be used.

  That is, first, the inventory transition of all the storage tanks is predicted and calculated within the prediction range of the future physical distribution state set for about 2 to 3 hours set by the condition setting and intake unit 30 in FIG. As a result of this prediction calculation, the storage tank that is predicted to cut the replenishment level of about 50% to 70% (can be set separately for each storage tank) set in the condition setting and intake section 30 in FIG. Extracted as candidates for storage tanks to be stored. In this example, the storage tanks 1, 2, and 3 are extracted as supply target tanks, and the storage tank 4 is regarded as not requiring supply at the current time and is excluded from the candidates.

  Next, the process path allocation pattern search unit 23 calculates the process path as shown in FIG. 6 for the extracted storage tanks to be replenished, and derives the entire process path allocation pattern.

  Details of the operation of the process path allocation pattern search unit 23 are shown below. The process path is calculated for each of the storage tanks extracted as the replenishment target, and the entire process path combination pattern is derived. First, the process route search information table 61 is fetched from the condition setting and fetching unit 30 in FIG. 3 based on the distribution structure, the yard / mountain arrangement, and the storage tank loading brand. For example, the case of the storage tank 2 is taken as an example. In STEP1, the storage tank 2 is searched from the starting facility of the process path search information table 61 (step S61).

  Next, in STEP 2, a brand that matches the brand B loaded in the storage tank 2 is searched from the mountain brands in the process path search information table 61 (step S62).

  In STEP 3, a set of yard and reclaimer corresponding to the searched mountain brand is searched (step S63). Here, it is understood that (yard 1, RR (reclaimer) No. 1) and (yard 2, RR No. 2) can be used.

  In STEP 4, a usable belt conveyor series is searched from the place where the searched starting equipment line and the searched mountain brand intersect (step S64). In this case, it is understood that the series 2 can be used when (yard 1, RR No. 1) is used, and the series 4 can be used when (yard 2, RR No. 2) is used.

In STEP5, an allocation pattern is calculated for the usable process paths guided to all supply target storage tanks (step S65). Here, storage tank 1 is (yard 1, RR No. 1, series 1), and storage tank 2 is (yard 1, RR No. 1, series 2), (yard 2, RR No. 2, series 4) ), Because storage tank 3 is (yard 1, RR No. 1, series 3), the total allocation pattern is
As allocation pattern 1,
(Reservoir 1, Yard 1, RR No.1, Series 1), (Reservoir 2, Yard 1, RR N o.1, Series 2), (Reservoir 3, Yard 1, RR No. 1) , Series 3)
And as allocation pattern 2,
(Storage tank 1, yard 1, RR No. 1, series 1), (Storage tank 2, yard 2, RR No. 2, series 4), (Storage tank 3, yard 1, RR No. 1) , Series 3)
These two patterns are derived.

  Next, a distribution model may be formulated for all the derived allocation patterns, here allocation patterns 1 and 2, based on respective setting conditions, distribution restrictions, and distribution status, or some logic is introduced. , Formulate a logistics model based on each setting condition, logistics constraints, and logistics status for only some of the patterns that are likely to be optimal solution candidates, and excluded allocation patterns are outside the scope of consideration It is also good.

  For example, as this logic, when there are many facilities or processes that interfere with each other in the detection pattern, they are excluded from the formulation. When this logic is applied to this example, in the allocation pattern 1, since all RR No. 1 is used for transportation to the storage tanks 1, 2, and 3, there are many interferences in RR No. 1, compared with that. Allocation pattern 2 uses RR No. 1 only for transportation to storage tanks 1 and 3.

  If logic that is targeted only when there are two or fewer interferences of equipment used for transportation to the same storage tank is introduced, the allocation pattern 1 is excluded from the candidate patterns, and only the allocation pattern 2 is formulated.

  Next, a distribution model is constructed by the distribution model construction unit 24 for all the allocation patterns of the obtained process paths. In this example, the case where formulation is performed for all the allocation patterns is taken as an example, and the concept of this formulation is shown in the concept of physical distribution model construction (FIG. 7).

  As shown in FIG. 7, it is constructed from an inter-process constraint model describing constraints between processes within one JOB and an inter-job constraint model modeling interference between JOBs.

In the inter-process constraint model, if the operation start time of the reclaimer is t _s , t _e , the transfer start time t_bc _s , t_bc _{e of the} conveyor belt series, and the tank entry start time is t_R _s , t_R _e , there is a certain time difference between the processes (L, m, n, and p are constants). The constraint in this case is
t_bc _s = t _s + l (1)
t_bc _e = t _e + m (2)
t_R _s = t _s + n (3)
t_R _e = t _e + p (4)
It is expressed.

In addition, if the tank stock level at the start of the storage tank is R (t _s ) and the tank stock level at the end of the tank is R (t _e ), the amount of tanks entering and leaving the storage tank is The constraint when it is constant regardless of time is
R (t _s ) = at _s + b (5)
R (t _e ) = ct _e + d (6)
It is expressed. Here, a, b, c, d are proportional constants representing the relationship between the time and the tank stock level.

In the inter-JOB constraint model, allocation pattern 2 (reservoir 1, yard 1, RR No. 1, series 1), (reservoir 2, yard 2, RR No. 2, series 4), (reservoir 3 In the case of Yard 1, RR No. 1, Series 3), JOB No. 1 is used for JOB (JOB 1) entering the storage tank 1 and JOB (JOB 3) entering the storage tank 3. It is necessary to use this equipment, but it is not possible to use it with time overlap (temporal interference). JOB 1 reclaimer No. 1 (RR No. 1) operation start time is t _s1 , operation end time is t _e1 , JOB 3 reclaimer No. 1 (RR No. 1) operation start time is t _s3 , operation end time is t _{Assuming e3} , the constraint in this case is
When JOB1 is processed earlier than JOB3 t _s3 ≧ t _e1 (7)
When JOB3 is processed earlier than JOB1 t _s1 ≧ t _e3 (8)
It is expressed.

Here, if a positive real number M and an integer I of 0 or 1 that are sufficiently larger than the time difference between the time at which JOB1 processing is performed and the time at which JOB2 processing is performed are introduced into the above equations (7) and (8), (7 ) And (8) can be expressed by the following formulas (9) and (10) which do not require case separation.
t _s3 -t _e1 + MI ≧ 0 ...... (9)
t _s1 -t _e3 + M (1-I) ≥ 0 (10)

Furthermore, when these equations are transformed, the logistics model becomes
AX ≤ B (11)
Xmin ≤ X ≤ Xmax (12)
∃x: integer for [x | x ∈ X] (13)
It can be constructed as a simple linear form and an integer constraint expression.

  X is the operation start / end time of each facility and storage tank inventory, I is a matrix representation, A and B are predetermined determinants, and Xmin and Xmax are the earliest and latest operation start times of each facility, respectively. In addition, the lower limit level and the upper limit level of the storage tank inventory level are expressed in a matrix, and the element of X corresponding to the integer constraint corresponding to Expression (13) is I (I is a subset of X).

  Optimizing by evaluating each of the distribution model formulas composed of linear and integer constraint formulas constructed by the logistics model construction unit 24 with an evaluation function in a linear or quadratic format set by the evaluation function setting unit 26 In the calculation unit 25, the optimal tank order, tank start and end time, tank amount, reclaimer operation start time, reclaimer operation end time as well as payout mountain, yard, used reclaimer, conveyor belt conveyor series, tank storage Calculate the mine tank. A commercially available solver for mixed integer programming may be used for the calculation of the optimum calculation unit 25.

  Here, the optimization calculation shown in FIG. 8 shows that the evaluation function that the smaller the difference from the tank entry target level 10% and the tank entry target level 90% is better, the calculation of this processing calculation unit is performed. The allocation pattern 2 having a small total value of all the storage tanks (a different weight may be assigned to each storage tank) of the difference between the entry / exit level and the target level is selected. As a result, tank 1 is discharged from yard 1, RR No. 1 operation start time is 17 minutes, operation end time is 47 minutes, belt conveyor series 1 transfer start time is 22 minutes, transfer end time is 52 minutes, input Tank start time is 27 minutes, tank entry end time is 57 minutes, tank 2 is discharged from yard 2, RR No. 2 operation start time is 18 minutes, operation end time is 48 minutes, belt conveyor series 4 transport The start time is 22 minutes, the transfer end time is 52 minutes, the entry tank start time is 27 minutes, the entry tank end time is 57 minutes, the tank 3 is dispensed from the yard 1, and the RR No. 1 operation start time is 41 Min, operation end time is 65 minutes, belt conveyor system 3 transfer start time is 41 minutes, transfer end time is 65 minutes, tank start time is 51 minutes, and tank end time is determined to be 75 minutes. .

  Among the optimal solutions calculated by the optimization unit 27, an appropriate number of JOBs is determined from the JOB with the earlier reclaimer operation start time, and the optimal tank entry order, tank start and end times are determined based on the determined information. The amount of tanks, the reclaimer operation start time, the reclaimer operation end time, as well as the payout mountain, the yard, the used reclaimer, the conveyor belt conveyor series, and the input tank storage tank are sent to the simulator 21.

  The simulator 21 advances the simulation time until the start of the reclaimer operation of the JOB with the latest reclaimer operation start time among the determined JOBs based on the given optimization calculation result.

  In this example, a case where two of the three storage tanks are determined is shown. In this case, tank 1 pays out from yard 1, RR No. 1 operation start time is 17 minutes, operation end time is 47 minutes, belt conveyor series 1 transfer start time is 22 minutes, transfer end time is 52 minutes, The tank entry time is 27 minutes, the tank entry time is 57 minutes, tank 2 is dispensed from Yard 2, RR No. 2 operation start time is 18 minutes, operation end time is 48 minutes, and belt conveyor system 4 is transported The start time is 22 minutes, the transfer end time is 52 minutes, the tank start time is 27 minutes, and the tank end time is 57 minutes. Moreover, the optimization result regarding the storage tank 3 is discarded. The simulator 21 proceeds with the simulation until the reclaimer operation start time 18 minutes in the storage tank 2 in accordance with the instruction determined for the storage tanks 1 and 2.

  Then, the simulation is performed again with the time advanced in this way, and the current distribution state is calculated and supplied to the optimization unit 27. In the same manner, the calculation of the simulator 21 and the optimization unit 27 is repeated every time the reclaimer operation start event of the JOB having the latest reclaimer operation start time among the JOBs determined by the previous optimization calculation occurs.

  If the schedule for all sections has been completed, the process ends. Otherwise, the process returns to the inventory transition prediction, and the entire schedule is created by repeating a series of processes after the inventory transition prediction as in the previous case.

  As described above, in this embodiment, the distribution model based on the current distribution state and distribution constraint rule obtained from the Petri net simulator 21 and the evaluation function for evaluating the distribution model are represented by a simple linear form and integer constraint expression. It is possible to create a logistics production / distribution plan at a high speed by applying mathematical programming to a Petri net model, and to realize analytical logistics optimization.

  Here, the simulator 21 of FIG. 1 is combined with the discrete event system simulator 310 shown in FIG. 10 by the inventory transition prediction unit 22 of the optimization unit 27, the process path allocation pattern search unit 23, the logistics model construction unit 24, and the like. 10, the optimization calculation unit 25 corresponds to the optimization calculation device 330 shown in FIG. 10, and the evaluation function setting unit 26 corresponds to the evaluation function setting device 340 shown in FIG. Note that the evaluation function setting unit 26 can selectively switch and set various types of evaluation functions.

  As described above in detail, according to the production / distribution plan creation apparatus of this embodiment, the distribution state that changes for each reclaimer operation start event of the JOB with the latest reclaimer operation start time among the confirmed JOBs A process model that is estimated by the simulator 21 based on the model, and that extracts a process path allocation pattern that changes for each reclaimer operation start event event of the JOB with the latest reclaimer operation start time among the determined jobs, and a logistics model that changes for each allocation pattern Is expressed in a simple linear format and integer constraint expression, and the evaluation function is expressed in a simple linear or quadratic format. The order of tanks, tank start and end times, tank input, reclaimer operation start time, reclaimer by mathematical programming Not only the operation end time, but also the payout mountain, yard, used reclaimer, conveyor belt conveyor series, and optimal storage tank storage tank Thereby achieving the (processing order, processing time, optimization of the treatment facility selection).

  That is, in this embodiment, a logistics model is constructed for each Petri net model of a manufacturing process and each process path allocation pattern, and a combination of mathematical programming methods is used to process multiple products in different multiple process paths, and each product. Even when multiple process paths having different values can be selected, the optimization calculation can be performed very easily by the mathematical programming method.

  In the above embodiment, the case where the present invention is applied to a production / distribution plan creation apparatus (simulator) has been described. However, the present invention can also be applied to a distribution control apparatus. In this case, an actual plant control device or the like is applied instead of the simulator 21 shown in FIG. In this way, the actual plant has a payout mountain as well as the optimal tank order, tank start and end times, tank amount, reclaimer operation start time, and reclaimer operation end time determined by the optimization unit 27. The material yard operation is executed according to the yard, the used reclaimer, the conveyor belt conveyor line, and the incoming tank storage tank.

  When one or a plurality of JOBs are performed in this way, the current distribution state in the actual plant changes, so that information is extracted and supplied to the optimum unit 27. In the optimization unit 27, the inventory transition prediction unit 22 performs inventory transition prediction, and extracts the storage tank to be replenished. The process path allocation pattern search unit 23 calculates the process path for the extracted replenishment target storage tank, and derives the entire process path allocation pattern. Here, with respect to the entire allocation pattern of the obtained process route, a distribution model consisting of a linear format and an integer constraint expression based on the given current distribution state and distribution constraints is applied to the entire process route allocation pattern, respectively. Build and supply to the optimization calculator 26. The optimization calculation unit 26 evaluates a given physical distribution model expression with an evaluation function, so that not only the order of tank entry, tank start and end times, tank amount, reclaimer operation start time, and reclaimer operation end time are provided. Calculate payout hills, yards, reclaimers used, conveyor belt conveyors, and tank storage tanks. Thereafter, the distribution control is executed by repeating the same processing.

  In the above embodiment, the Petri net model is given as an example of the simulator 21 model. However, the present invention is not limited to this, and a model that can output current physical condition information for each event. Any simulator can be applied.

  The simulator 21 and the optimization unit 27 are configured by a microcomputer including a CPU (Central Processing Unit), a RAM (Random Access Memory), a ROM (Read Only Memory), and the like, for example, a personal computer or the like. It can be realized by a computer.

It is a block diagram which shows the principal part structure of the production and distribution plan preparation apparatus of this embodiment. It is a schematic diagram of a raw material yard manufacturing process (conveyance). It is a figure which shows the positioning of the production and distribution plan preparation apparatus by this embodiment. It is a figure which shows the simple example which reduced the scale of the raw material yard manufacturing process (conveyance) used in order to demonstrate the outline | summary of a process. It is a figure for demonstrating storage tank stock forecast transition. It is a flowchart for explaining a process route allocation pattern search and a flowchart showing a search method. It is a figure for demonstrating the content which represented the distribution restrictions by the linear format and the integer restrictions. It is a figure for demonstrating the outline | summary of the method of extracting the optimal thing from the physical distribution model constructed | assembled for every process allocation pattern. It is a Petri net expression diagram of a simple example in which the scale of the raw material yard manufacturing process (conveyance) used for explaining the outline of the process is the process outline of the production / distribution plan creation unit according to the present embodiment. It is a block diagram explaining an example of the whole structure of the production and physical distribution plan preparation apparatus to which this invention is applied.

Explanation of symbols

21 Simulator by Petri Net Model 22 Prediction Unit for Reservoir Storage Inventory 23 Process Route Allocation Pattern Retrieval Unit 24 Logistics Model Construction Unit 25 Optimization Calculation Unit 26 Evaluation Function Setting Unit 27 Optimization Unit 30 Condition Setting and Capture Unit 31 Production / Logistics Plan creation section 32 Gantt chart display / storage tank inventory transition graph display section 33 Various evaluation sections

Claims (9)

To produce a production and / or logistics plan in a production and / or logistics process in which a plurality of production and / or logistics objects are processed in different multi-process paths, and each object can select a different multi-process path. Production and / or logistics planning equipment
The production and / or logistics flow and production and / or logistics constraints in actual operation of the production and / or logistics process are represented by a model, and the production and / or logistics status and production and / or logistics constraints when an event occurs are represented. A simulator to detect,
Production and / or distribution status and production and / or distribution restrictions of the production and / or distribution process for a target period (plan creation period) set in advance from the production start date and time of the production and / or distribution plan. Among them, a mathematical model holding device that stores a mathematical model consisting only of a preset part,
An evaluation function setting device for setting an evaluation function for evaluating the mathematical model held by the mathematical model holding device;
An optimization calculation device that performs optimization calculation processing using the mathematical formula model held by the mathematical formula model holding device and the evaluation function set by the evaluation function setting device to calculate a physical distribution instruction for the simulator; And
Through the optimization calculation processing by the optimization calculation device, the distribution instruction is calculated for the target period (instruction calculation period) set within a predetermined range from the planning start date and time, is given to the simulator, and is selected in advance. The simulation is executed for the target period (simulation period), the simulation result is confirmed as a production and / or distribution plan for a preset period (plan decision period), and the date and time immediately after the determined period is newly set. production and / or distribution to create a production and / or logistics planning from the simulation results obtained by repeating the process of devising the production and / or logistics planning and set as planning start date in the production and / or logistics process A planning device,
The formula model is constructed for all patterns or a part of patterns of combination processes that can be selected as the process path, and an optimization calculation process is performed by the optimization calculation device for each of the constructed formula models. A production and / or distribution plan creation device characterized by selecting one. The production and / or distribution plan creation device according to claim 1, wherein the evaluation function is linear or quadratic. 3. The production and / or distribution plan creation device according to claim 1, wherein the model is a Petri net model. The production and / or distribution plan creation device according to claim 1, wherein the optimization calculation device performs optimization calculation processing as an optimization or quasi-optimization problem. To produce a production and / or logistics plan in a production and / or logistics process in which a plurality of production and / or logistics objects are processed in different multi-process paths, and each object can select a different multi-process path. In the production and / or logistics planning method of
The production and / or logistics flow and production and / or logistics constraints in actual operation of the production and / or logistics process are represented by a model, and the production and / or logistics status and production and / or logistics constraints when an event occurs are represented. A simulator to detect,
Production and / or distribution status and production and / or distribution restrictions of the production and / or distribution process for a target period (plan creation period) set in advance from the production start date and time of the production and / or distribution plan. Among them, a mathematical model holding device that stores a mathematical model consisting only of a preset part,
An evaluation function setting device for setting an evaluation function for evaluating the mathematical model held by the mathematical model holding device;
Using an optimization calculation apparatus that performs optimization calculation processing using the mathematical expression model held by the mathematical expression model holding apparatus and the evaluation function set by the evaluation function setting apparatus to calculate a physical distribution instruction for the simulator ,
Through the optimization calculation processing by the optimization calculation device, the distribution instruction is calculated for the target period (instruction calculation period) set within a predetermined range from the planning start date and time, is given to the simulator, and is selected in advance. The simulation is executed for the target period (simulation period), the simulation result is confirmed as a production and / or distribution plan for a preset period (plan decision period), and the date and time immediately after the determined period is newly set. production and / or distribution to create a production and / or logistics planning from the simulation results obtained by repeating the process of devising the production and / or logistics planning and set as planning start date in the production and / or logistics process A plan creation method,
The formula model is constructed for all patterns or a part of patterns of combination processes that can be selected as the process path, and an optimization calculation process is performed by the optimization calculation device for each of the constructed formula models. A production and / or distribution plan creation method characterized by selecting one. To produce a production and / or logistics plan in a production and / or logistics process in which a plurality of production and / or logistics objects are processed in different multi-process paths, and each object can select a different multi-process path. In a computer program for causing a computer to execute production and / or distribution planning of
The production and / or logistics flow and production and / or logistics constraints in actual operation of the production and / or logistics process are represented by a model, and the production and / or logistics status and production and / or logistics constraints when an event occurs are represented. A simulator to detect,
Production and / or distribution status and production and / or distribution restrictions of the production and / or distribution process for a target period (plan creation period) set in advance from the production start date and time of the production and / or distribution plan. Among them, a mathematical model holding device that stores a mathematical model consisting only of a preset part,
An evaluation function setting device for setting an evaluation function for evaluating the mathematical model held by the mathematical model holding device;
Using an optimization calculation apparatus that performs optimization calculation processing using the mathematical expression model held by the mathematical expression model holding apparatus and the evaluation function set by the evaluation function setting apparatus to calculate a physical distribution instruction for the simulator ,
Through the optimization calculation processing by the optimization calculation device, the distribution instruction is calculated for the target period (instruction calculation period) set within a predetermined range from the planning start date and time, is given to the simulator, and is selected in advance. The simulation is executed for the target period (simulation period), the simulation result is confirmed as a production and / or distribution plan for a preset period (plan decision period), and the date and time immediately after the determined period is newly set. production and / or distribution to create a production and / or logistics planning from the simulation results obtained by repeating the process of devising the production and / or logistics planning and set as planning start date in the production and / or logistics process A computer program that causes a computer to execute planning,
The formula model is constructed for all patterns or a part of patterns of combination processes that can be selected as the process path, and an optimization calculation process is performed by the optimization calculation device for each of the constructed formula models. A computer program characterized by selecting things. In a process control apparatus for controlling a production and / or logistics process in which a plurality of production and / or logistics objects are processed in different multi-step paths, and each object can select a different multi-step path,
The production and / or logistics flow and production and / or logistics constraints in actual operation of the production and / or logistics process are represented by a model, and the production and / or logistics status and production and / or logistics constraints when an event occurs are represented. A simulator to detect,
Production and / or distribution status and production and / or distribution restrictions of the production and / or distribution process for a target period (plan creation period) set in advance from the production start date and time of the production and / or distribution plan. Among them, a mathematical model holding device that stores a mathematical model consisting only of a preset part,
An evaluation function setting device for setting an evaluation function for evaluating the mathematical model held by the mathematical model holding device;
An optimization calculation device that performs optimization calculation processing using the mathematical formula model held by the mathematical formula model holding device and the evaluation function set by the evaluation function setting device to calculate a physical distribution instruction for the simulator; And
Through the optimization calculation processing by the optimization calculation device, the distribution instruction is calculated for the target period (instruction calculation period) set within a predetermined range from the planning start date and time, is given to the simulator, and is selected in advance. The simulation is executed for the target period (simulation period), the simulation result is confirmed as a production and / or distribution plan for a preset period (plan decision period), and the date and time immediately after the determined period is newly set. A process control device for controlling the production and / or distribution process based on a simulation result obtained by repeating a process of setting a planning start date and time and planning a production and / or distribution plan,
The formula model is constructed for all patterns or a part of patterns of combination processes that can be selected as the process path, and an optimization calculation process is performed by the optimization calculation device for each of the constructed formula models. A process control device characterized by selecting one. In a process control method for controlling a production and / or distribution process in which a plurality of production and / or distribution objects are processed in different multi-step paths, and each object can select a different multi-step path,
The production and / or logistics flow and production and / or logistics constraints in actual operation of the production and / or logistics process are represented by a model, and the production and / or logistics status and production and / or logistics constraints when an event occurs are represented. A simulator to detect,
Production and / or distribution status and production and / or distribution restrictions of the production and / or distribution process for a target period (plan creation period) set in advance from the production start date and time of the production and / or distribution plan. Among them, a mathematical model holding device that stores a mathematical model consisting only of a preset part,
An evaluation function setting device for setting an evaluation function for evaluating the mathematical model held by the mathematical model holding device;
Using an optimization calculation apparatus that performs optimization calculation processing using the mathematical expression model held by the mathematical expression model holding apparatus and the evaluation function set by the evaluation function setting apparatus to calculate a physical distribution instruction for the simulator ,
Through the optimization calculation processing by the optimization calculation device, the distribution instruction is calculated for the target period (instruction calculation period) set within a predetermined range from the planning start date and time, is given to the simulator, and is selected in advance. The simulation is executed for the target period (simulation period), the simulation result is confirmed as a production and / or distribution plan for a preset period (plan decision period), and the date and time immediately after the determined period is newly set. A process control method for controlling the production and / or logistics process based on a simulation result obtained by repeating a process of setting a planning start date and time and planning a production and / or logistics plan,
The formula model is constructed for all patterns or a part of patterns of combination processes that can be selected as the process path, and an optimization calculation process is performed by the optimization calculation device for each of the constructed formula models. A process control method characterized by selecting one. Process control to control production and / or logistics processes that can handle multiple production and / or logistics objects with different multi-step paths and each object can select different multi-step paths Computer program
The production and / or logistics flow and production and / or logistics constraints in actual operation of the production and / or logistics process are represented by a model, and the production and / or logistics status and production and / or logistics constraints when an event occurs are represented. A simulator to detect,
Production and / or distribution status and production and / or distribution restrictions of the production and / or distribution process for a target period (plan creation period) set in advance from the production start date and time of the production and / or distribution plan. Among them, a mathematical model holding device that stores a mathematical model consisting only of a preset part,
An evaluation function setting device for setting an evaluation function for evaluating the mathematical model held by the mathematical model holding device, a mathematical model held by the mathematical model holding device, and an evaluation set by the evaluation function setting device Using an optimization calculation device that performs optimization calculation processing using a function and calculates logistics instructions for the simulator,
Through the optimization calculation processing by the optimization calculation device, the distribution instruction is calculated for the target period (instruction calculation period) set within a predetermined range from the planning start date and time, is given to the simulator, and is selected in advance. The simulation is executed for the target period (simulation period), the simulation result is confirmed as a production and / or distribution plan for a preset period (plan decision period), and the date and time immediately after the determined period is newly set. A computer program for causing a computer to execute control of the production and / or distribution process based on a simulation result obtained by repeating a process of setting a planning start date and time and planning a production and / or distribution plan,
The formula model is constructed for all patterns or a part of patterns of combination processes that can be selected as the process path, and an optimization calculation process is performed by the optimization calculation device for each of the constructed formula models. A computer program characterized by selecting things.

Images