Stacking method, stacking device, and stacking program

Abstract

To provide a technique capable of obtaining a good accuracy solution within a practical time when a stacking plan is created for a ship transporting cargoes and ballast water.SOLUTION: A stacking method according to the present invention divides each of a hold and a ballast tank into two types of a large slice and a small slice and determines a loading order of cargoes by evaluating a loading state using the maximum cargo quantity constraint of the cargoes and the minimum cargo quantity constraint of the cargoes.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a stowage method for determining the order in which cargo is stowed onto a ship. [Background technology]

[0002] Ships that transport cargo are equipped with one or more containers (holds) for storing cargo and one or more ballast tanks for storing ballast water. When loading cargo onto a ship, the order in which it is loaded determines the efficiency of the work. This effect occurs both at the loading port and at the unloading port. For example, if the cargo to be unloaded at the loading port is placed deep within the hold, the efficiency of the work decreases. Or, it is considered inefficient to remove cargo from a lower hold when there is cargo in the upper hold.

[0003] The following Patent Document 1 describes a technology that aims to "automatically create the best stacking position plan under any conditions, and can also handle the case where the work berth is shifted during the stacking work," and that "includes an arrangement plan creation process that creates an appropriate stacking arrangement plan and calculates a predetermined evaluation value, a changed arrangement plan creation process that creates a modified arrangement plan of the stacking arrangement plan and calculates a predetermined evaluation value, an arrangement plan selection process that selects a better arrangement plan based on the comparison result of the two evaluation values, and a stacking position optimization process that repeats the arrangement plan creation process, the changed arrangement plan creation process, and the arrangement plan selection process, with the selected arrangement plan or a new arrangement plan being the stacking arrangement plan in the arrangement plan creation process. Furthermore, when the work berth is changed during stacking, a work berth allocation process that allocates each product to each work berth based on the work berth allocation information and predetermined rules" (see abstract). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 10-305929 Summary of the Invention [Problem to be solved by the invention]

[0005] In order to optimally allocate the cargo carried in the holds and the ballast water carried in the ballast tanks, it is possible to formulate the loading of these as a combinatorial optimization problem and obtain the optimal solution using an optimization algorithm such as simulated annealing. Specifically, for example, the holds and ballast tanks can be divided into slices of equal capacity, and the cargo and ballast water can also be divided into slices of equal capacity, and the combination of these slices can be obtained.

[0006] However, for example, if a ship with a displacement of 200,000 tons is used, the capacity of one slice is 1000m 3 Even if we set it to 1,000m, the number of state variables representing the state of each slice would be tens of thousands, and the memory capacity consumed when executing the optimization algorithm would be too large to be practical. 3 However, slicing by this method may not be accurate enough, and some data may be left behind.

[0007] Conventional loading methods such as those in Patent Document 1 do not provide a practical method for solving combinatorial optimization problems involving a huge number of state variables. Therefore, even if the method in this document were applied as is to the above example, it would be difficult to obtain a solution with good accuracy within a practical time frame.

[0008] The present invention has been made in consideration of the above-mentioned problems, and has an object to provide a technique that can obtain an accurate solution within a practical time frame when creating a stowage plan for a ship carrying cargo and ballast water. [Means for solving the problem]

[0009] The loading method of the present invention divides holds and ballast tanks into two types of slices, large slices and small slices, and determines the loading order of cargo by evaluating the loading condition using maximum cargo quantity constraints and minimum cargo quantity constraints. Effect of the Invention

[0010] According to the stowage method of the present invention, when creating a stowage plan for a ship that transports cargo and ballast water, a highly accurate solution can be obtained within a practical time. Problems, configurations, advantages, etc. other than those described above will become apparent from the description of the following embodiments. [Brief description of the drawings]

[0011] [Figure 1] FIG. 2 is a diagram showing a schematic structure of a hold and a ballast tank provided on the ship according to the first embodiment. [Diagram 2] FIG. 2 is a diagram for explaining the definition of an evaluation function H. [Figure 3A] An example of state variable q is shown below. [Figure 3B] FIG. 3B is an enlarged view of the area surrounded by the dotted line in FIG. 3A. [Figure 4] FIG. 13 is a diagram for explaining a loading evaluation coefficient among Cij. [Diagram 5] FIG. 13 is a diagram for explaining a loading / unloading evaluation coefficient among Cij. [Figure 6] FIG. 13 is a diagram for explaining an example of CTij. [Figure 7A] FIG. 13 is a diagram showing an example of an Ank. [Figure 7B] FIG. 13 is a conceptual diagram showing an example of an Ank in which holds are arranged according to the order specified by the Loading Sequence. [Figure 8] FIG. 13 is a diagram for explaining a maximum quantity constraint in the penalty function HA. [Figure 9] FIG. 13 is a diagram for explaining a minimum quantity constraint in the penalty function HA. [Figure 10] FIG. 13 is a diagram for explaining a hold capacitance constraint in the penalty function HA. [Figure 11] FIG. 13 is a diagram illustrating a hold weight constraint in the penalty function HA. [Figure 12] FIG. 13 is a diagram illustrating adjacent hold constraints. [Figure 13] 1 is a graph illustrating an example of a maximum allowable weight and a minimum required weight. [Figure 14] FIG. 13 is a diagram for explaining the draft constraint of the penalty function HA. [Figure 15] FIG. 13 is a diagram for explaining a trim constraint in the penalty function HA. [Figure 16] FIG. 13 is a diagram for explaining a frame position where a vertical intensity constraint is imposed in the penalty function HA. [Figure 17] FIG. 13 is a diagram for explaining details of the vertical strength constraint. [Figure 18] FIG. 1 is a diagram illustrating a procedure for calculating shear stress SS(n) and bending moment SB(n). [Figure 19] FIG. 13 is a process flow diagram illustrating a process of searching for an optimal solution for a state variable q. [Figure 20] FIG. 13 is a schematic diagram illustrating surface layer sorting. [Figure 21] FIG. 11 is a configuration diagram of a stowage apparatus 1 according to a second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] <Embodiment 1> As mentioned above, when holds and ballast tanks (as well as cargo and ballast water) are divided into slices of equal volume, the number of state variables reaches tens of thousands, making it difficult to obtain an accurate combinatorial optimal solution within a practical time frame. The number of variables increases further because variables are required not only in the cost function but also in the constraint functions. Therefore, it is desirable to keep the number of state variables below 10,000. On the other hand, it is necessary to ensure that stowage accuracy does not decrease significantly by limiting the number of state variables.

[0013] Therefore, in the embodiment of the present invention, the holds and ballast tanks are divided into two types, a large-capacity slice and a small-capacity slice. For example, all of the holds and ballast tanks are divided into 1000m 3 Instead of splitting it into slices of 1000m, the smaller slices are 1000m 3 , large capacity slices are 10,000m 3 This aims to significantly reduce the total number of slices.

[0014] FIG. 1 is a diagram showing a schematic structure of holds and ballast tanks provided on a ship according to a first embodiment of the present invention. In FIG. 1, there are nine holds and eight ballast tanks. Even if the capacity of each hold or each ballast tank is different, the number of slices is the same. For example, all holds are divided into two large hold slices and six small hold slices, and all ballast tanks are divided into one large ballast tank slice and seven small ballast tank slices. This ensures that the conditions are the same regardless of which holds or ballast tanks are loaded with cargo or ballast water.

[0015] The cargo to be loaded into the holds and the ballast water to be loaded into the ballast tanks are similarly divided into large and small slices. The large cargo slices are loaded into the large hold slices, and the small cargo slices are loaded into the small hold slices. Similarly, the large ballast water slices are loaded into the large ballast tank slices, and the small ballast water slices are loaded into the small ballast tank slices.

[0016] FIG. 2 is a diagram for explaining the definition of the evaluation function H. The evaluation function H is a penalty function H A and the connection energy function H B The connection energy function H B is to solve the optimal luggage loading state by a search algorithm by formulating the luggage slice loading state as the connection energy between luggage slices and the connection energy between hold slices, and searching the state variables to minimize the connection energy. The penalty function H A More on this later.

[0017] In the formula in Fig. 2, q is a state variable that represents the position of the baggage slice. C ij is a luggage slice evaluation coefficient used to evaluate the layout relationship between luggage slices by weighting the connection energy between the luggage slices. A nk is a hold slice evaluation coefficient used to evaluate the arrangement relationship between the hold slices by weighting the connection energy between the hold slices.

[0018] In this embodiment, a state variable q that minimizes the evaluation function H (minimizes the connection energy) is searched for by an optimization algorithm such as simulated annealing.

[0019] FIG. 3A shows an example of a state variable q. When baggage j is loaded in hold i, the state variable q is q ij =1, when not installed q ij = 0, and the state variable q is a matrix with elements of (number of hold slices + number of ballast tank slices) × (number of cargo slices + number of ballast water slices). The state variable q has maximum and minimum quantity constraints shown in the upper right of Figure 3. These constraints are imposed, for example, by contract quantities. The dotted line parts will be described later.

[0020] The minimum quantity constraint specifies the minimum number of cargo slices that should be loaded onto a large hold slice and the minimum number of cargo slices that should be loaded onto a small hold slice for each type of cargo (Type#0 to Type#2 in this example). For example, for Type#0 cargo, a minimum of five pieces must be loaded onto a large hold slice and a minimum of three pieces must be loaded onto a small hold slice. Ballast water and ballast tank slices have similar constraints, but there is only one type of ballast water (BW).

[0021] The maximum quantity constraint specifies the maximum number of cargo slices that should be loaded onto a large hold slice and the maximum number of cargo slices that should be loaded onto a small hold slice for each type of cargo (Type#0 to Type#2). For example, for Type#0 cargo, a maximum of eight pieces can be loaded onto a large hold slice and a maximum of five pieces onto a small hold slice. Ballast water and ballast tank slices have similar constraints, but there is only one type of ballast water (BW).

[0022] The state variable q when loading cargo to satisfy the above constraints is configured as shown in the left diagram of Figure 3. The shaded areas indicate that no cargo is loaded. The cargo slices indicated by the black areas must be loaded onto one of the hold slices (minimum quantity constraint). The cargo slices indicated by the white areas may or may not be loaded onto the hold slice (maximum quantity constraint).

[0023] The total number of slices must meet the following rules: The total number of large hold slices must be equal to or greater than the minimum quantity of cargo to be loaded into large hold slices, combined for all cargo types. The total number of small hold slices must be equal to or greater than the minimum quantity of cargo to be loaded into small hold slices, combined for all cargo types. The total number of large ballast tank slices must be equal to or greater than the minimum quantity of ballast water to be loaded into large ballast tank slices. The total number of small ballast tank slices must be equal to or greater than the minimum quantity of ballast water to be loaded into small ballast tank slices.

[0024] Ballast water may be loaded not only in ballast tanks, but also in even-numbered holds, for example. For example, if the cargo is very light, the stability of the ship will be poor, so ballast water is loaded in ballast tanks, but if that is still not enough, ballast water is also loaded in even-numbered holds. In this case, in addition to the number of ballast tank slices, the number of even-numbered hold slices will also be counted as slices that can load ballast water. Therefore, the rule for the total number of slices is as follows: The sum of the total number of large ballast tank slices and the total number of even-numbered large hold slices must be equal to or greater than the minimum number of ballast water slices to be loaded in a large ballast tank slice. The sum of the total number of small ballast tank slices and the total number of even-numbered small hold slices must be equal to or greater than the minimum number of ballast water slices to be loaded in a small ballast tank slice.

[0025] Fig. 3B is an enlarged view of the portion enclosed by the dotted line in Fig. 3A. Examples of maximum quantity constraints and minimum quantity constraints will be described with reference to Fig. 3B.

[0026] The minimum quantity constraint is explained below. The minimum quantity constraint is represented by the black cells. In this example, a minimum of five large baggage slices of Type#0 must be loaded. These five large baggage slices are always loaded into one of the black cells surrounded by dotted lines. For example, when the first large baggage slice is loaded into the first large hold slice of CH#0, a 1 is set in the top left cell in FIG. 3B. All the cells above and below it are set to 0 (i.e., the same large baggage slice will not be loaded onto another large hold slice). Since the first large hold slice of CH#0 has already been used, all the cells to the left and right of it are set to 0. The second and subsequent large baggage slices are loaded in the same way, and all the cells above, below, left and right of the black cell set to 1 are set to 0.

[0027] The maximum quantity constraint will now be explained. The maximum quantity constraint is represented by the blank space. In this example, a maximum of eight Type#0 large baggage slices can be loaded. Since a minimum of five large baggage slices have already been loaded, the remaining three large baggage slices may or may not be loaded. For example, when a sixth large baggage slice is loaded into the second large hold slice of CH#1, a 1 is set in the blank space, as shown in the third column from the right in Figure 3B. All the spaces above, below, to the left and right of the slice are set to 0. The seventh and eighth large baggage slices are loaded in the same way. The sixth to eighth large baggage slices do not have to be loaded, but if they are loaded, they should conform to the above example.

[0028] As shown in the above examples, maximum quantity constraints and minimum quantity constraints can be expressed as follows. For black cells representing minimum quantity constraints, the sum of the state variables belonging to the same column will be 1. For white cells representing maximum quantity constraints, it is okay to set them to 1 or not, but if they are set, the cells above, below, left and right will be 0. In other words, for white cells, the sum of the state variables belonging to the same row will be 0 or 1, and the sum of the state variables belonging to the same column will also be 0 or 1.

[0029] The maximum quantity constraint and the minimum quantity constraint can be explained as follows. First, the maximum quantity constraint and the minimum quantity constraint have in common the state variable q ij The sum of the state variables belonging to the same column is either 0 or 1. Furthermore, for minimum quantity constraints, the sum of the state variables belonging to the same column is 1.

[0030] Figure 4 shows the C ij FIG. 3 is a diagram illustrating the loading evaluation coefficient. ij is C ij =CL ij ×AMP LOADING +CD ij ×AMP DISCHARGE +CT ij It can be formulated as follows: CL ij ×AMP LOADING is the loading evaluation coefficient that evaluates the order in which the baggage slices are loaded. CDij ×AMP DISCHARGE is the loading and unloading evaluation coefficient for evaluating the order of loading and unloading the cargo slices. CT ij is the cargo type evaluation coefficient for evaluating whether the same cargo type is loaded on the bow side or lower layer of a certain cargo slice. Figure 4 shows an example of CL among these ij . AMP LOADING and AMP DISCHARGE can be set arbitrarily.

[0031] (1) Evaluating CL ij as -1 is in the following loading state: The loading port of the lower layer or bow side cargo j is in front of the loading port of the upper layer or stern side cargo i, that is, if the loading port of i > the loading port of j: CL[i, j] = -1 (2) Evaluating CL ij as +1 is in the loading state other than the above: The loading port of the lower layer or bow side cargo j is behind the loading port of the upper layer or stern side cargo i, that is, if the loading port of i < the loading port of j: CL[i, j] = 1 (3) Evaluating CL ij as 0 is in the loading state other than the above: The loading port of the lower layer or bow side cargo j is the same as the loading port of the upper layer or stern side cargo i, that is, if the loading port of i = the loading port of j: CL[i, j] = 0

[0032] Figure 5 is a diagram for explaining the loading and unloading evaluation coefficient among C ij . CD ij can be defined as follows.

[0033] (1) Evaluating CD ij as -1 is in the following loading state: The unloading port of the lower layer or bow side cargo j is behind the unloading port of the upper layer or stern side cargo i, that is, if the unloading port of i < the unloading port of j: CD[i, j] = -1 (2) Evaluating CD ij as +1 is in the loading state other than the above: The port of disembarkation of the lower or bow cargo j is earlier than the port of disembarkation of the upper or stern cargo i, i.e., if port of disembarkation of i>port of disembarkation of j: CD[i,j]=1 (3)CD ij is evaluated as 0 for all other loading conditions: The port of loading and disembarking of the lower or bow cargo j is the same as the port of loading and disembarking of the upper or stern cargo i, i.e., if port of loading and disembarking of i = port of loading and disembarking of j: CD[i,j] = 0

[0034] Figure 6 shows the CT ij FIG. 1 is a diagram illustrating an example of CT when a certain baggage slice and a baggage slice loaded on the bow side or the lower level are of the same baggage type. ij =-1 (improve the evaluation value). If the types of luggage are different, CT ij =AMP_TYPE (deteriorates the evaluation value). AMP_TYPE can be set arbitrarily, for example +1. For ballast water, CT ij The value of is not set (set to 0). In other words, unlike with cargo, the accumulation of the same type of ballast water slices in the same ballast tank is not evaluated for ballast water. Otherwise, ballast water would tend to accumulate in the ballast tanks. Note that the cargo slices that are left behind are calculated using the q ij = 0, so it is not included in the cost calculation.

[0035] FIG. 7A shows nk FIG. 1 is a diagram showing an example of nk q luggage slice ni andq kj It is used to evaluate whether the arrangement relationship between holds n and k is desirable when the nk is configured to increase the evaluation function H (deteriorate the evaluation value) when the placement relationship between holds n and k is undesirable.

[0036] If holds n and k are the same, A nk=AMP_INTRA=m (m is a positive integer of 2 or more). This is to strictly evaluate the mixing of different types of cargo in the same hold. However, unlike cargo, ballast water may be loaded in the same ballast tank. In other words, AMP_INTRA is not set for ballast tanks.

[0037] If holds n and k are arranged in such a way that cargo is loaded into the bow hold after being loaded into the stern hold, the evaluation function H is increased (the evaluation value is worsened). In this embodiment, there is only one level of holds and there is no vertical arrangement, so there is no weighting in the vertical direction.

[0038] Depending on the ship, the loading sequence for each hold may be specified in the Loading Manual. nk You may take this order into account when creating A. nk When creating the Loading Sequence, the holds are arranged from the bow in the order of the Loading Sequence described in the Loading Manual of the individual ship. For example, if the Loading Sequence indicates that cargo should be loaded in the order of hold numbers 719356284, then A nk The n rows of holds are arranged in the order of 719356284, starting from the row of ballast water tanks. This allows the arrangement relationship between holds n and k to be evaluated according to the Loading Sequence. Figure 7A shows an example in which the order specified by the Loading Sequence matches the actual order of the hold numbers.

[0039] Figure 7B shows the A-series data with holds arranged in the order specified by the Loading Sequence. nk This is a conceptual diagram showing an example of loading luggage into each hold in the order of the above example (hold number: 719356284). nk Hold number 4 is located at the bow end of the ship. nkIt is located at the sternmost position of the ship.

[0040] Figure 8 shows the penalty function H A FIG. 1 is a diagram illustrating the maximum quantity constraint among the penalty functions H A is a function constructed to impose a penalty on the evaluation function H when the state variable q does not satisfy the constraint. The penalty function H A is the sum of (a) cargo type contract quantity constraints, (b) hold capacity constraints, (c) hold weight constraints, (d) draft and trim constraints, and (e) longitudinal strength constraints.

[0041] In terms of the maximum quantity constraint, q ij must be a matrix in which the number of variables that are 1 in each row / column is one or less. The reason is as explained in Figures 3A and 3B. If the formula in Figure 8 is not satisfied, a penalty is added to the evaluation function H.

[0042] Figure 9 shows the penalty function H A From the viewpoint of the minimum quantity constraint, q ij must be a matrix whose sum is 1 in the parcel slice columns for the minimum contract quantity of each parcel type. The reason is as explained in Figures 3A and 3B. If the formula in Figure 9 cannot be satisfied, a penalty is added to the evaluation function H.

[0043] Figure 10 shows the penalty function H A This is a diagram explaining the hold capacity constraint. The total volume of baggage in the holds must not exceed the loadable capacity of the holds. This constraint must be satisfied for all holds. If the total volume of baggage in any one hold exceeds the loadable capacity of the hold, a penalty is added to the evaluation function H. This constraint is expressed by the formula in FIG.

[0044] Figure 11 shows the penalty function H AFIG. 11 is an explanatory diagram of the hold weight constraint. The total weight of baggage in a hold must not exceed the maximum allowable weight (Allowed_Mass) that the hold can carry, and must be equal to or greater than the minimum required weight (Required_Mass) that must be carried. This constraint must be satisfied for all holds. If the total baggage weight in any one hold does not satisfy the hold weight constraint, a penalty is added to the evaluation function H. This constraint is expressed by the formula in FIG. 11. Details of the maximum allowable weight and minimum required weight will be described later.

[0045] FIG. 12 is a diagram explaining adjacent hold constraints. There are single hold constraints and adjacent hold constraints. The adjacent hold constraint is a constraint that determines the maximum allowable weight and minimum required weight for the total luggage weight of two adjacent holds separated by a bulkhead (a partition between holds) as shown in FIG. 12. For example, in the case where adjacent constraints 1&2 and 2&3 exist, when luggage is added to hold 2, there is a high possibility that these will violate the constraints at the same time. In other words, the penalty value for violation of the adjacent hold constraint is doubled, and the weight of the constraint condition is rapidly doubled. This rapid increase in the weight of the constraint condition hinders the minimization of the objective function, and there is a strong tendency for a solution not to be obtained. Therefore, the adjacent hold weight constraint is set to 0.5, while the individual weights of the other constraint conditions are 1.0.

[0046] FIG. 13 is a graph showing an example of the allowable maximum weight and the required minimum weight. These values ​​are given by a linear graph like FIG. 13. This graph is given, for example, in the Loading Manual. This graph shows the allowable maximum weight and the required minimum weight for each value of the horizontal axis (draft) at the midpoint of each hold for single hold constraints, and shows the allowable maximum weight and the required minimum weight for each value of the horizontal axis (draft) at the midpoint of adjacent holds for adjacent hold constraints. Therefore, by referring to this graph using the draft, the allowable maximum weight and the required minimum weight can be obtained.

[0047] However, in this embodiment, in order to calculate the draft, trim and longitudinal center of gravity (LCG) are required, and both are calculated using the state variable q. In other words, in the process of searching for the state variable q, it is necessary to incorporate the allowable maximum weight and the required minimum weight as constraint functions, but the state variable q is required to calculate the allowable maximum weight and the required minimum weight, and there is a relationship in which the solution (q) and its constraint function change simultaneously. It is not realistic to solve such a multiple optimization problem.

[0048] Therefore, in this embodiment, it is assumed that the draft reaches DCF (draft equivalent to displacement) and the trim reaches maximum trim (TRIM_MAX) ± accuracy correction value when the optimal solution is reached, and each draft Dmh[fp] is calculated using formula 1 in Fig. 13. Then, the formula for calculating the allowable maximum weight and the required minimum weight from the graph in Fig. 13 and Dmp[fp] is formula (2) in Fig. 13. Therefore, in this embodiment, the allowable maximum weight and the required minimum weight are calculated using formula (2).

[0049] Figure 14 shows the penalty function H A This is a diagram explaining the draft constraint. For the draft constraint, the displacement Δ corresponding to the specified draft (DRAFT) is obtained from, for example, a hydrostatic table, and a solution is obtained so that the total weight DW (Dead Weight) of the ship approaches Δ as close as possible. For the trim constraint described later, the LCB (Center of longitudinal buoyancy) when DW=Δ is used, assuming that DW has sufficiently converged to the vicinity of Δ. Therefore, if DW has not sufficiently converged to the vicinity of Δ, an error will occur in optimization due to the trim constraint. Therefore, in the search process, the goal is to load cargo up to DW=0.999Δ. A penalty is added to state variables q that do not satisfy the draft constraint.

[0050] Figure 15 shows the penalty function H AThis is a diagram explaining the trim constraints. The center of buoyancy, trim moment, etc. at the even keel corresponding to the specified draft are obtained from the Hydrostatic Table, and the arm length from the midship of each element of the ship is obtained from the Tables for Capacity and Center of Gravity, and then the loading is determined so that the total rotational moment brings the attitude within the specified trim range. A penalty is added to state variables q that do not satisfy the trim constraints.

[0051] Figure 16 shows the penalty function H A This is a diagram explaining the frame positions for imposing vertical strength constraints. The shear stress and bending moment are calculated for each frame position, and the vertical strength constraint is that these are within the allowable values ​​at all frame positions. For FR26.5(0) in Figure 16, the components considered in the strength calculation are q ij is not included, it is added to the shear stress and bending moment as a constant, for example.

[0052] Figure 17 is a diagram for explaining the details of the longitudinal strength constraint. The longitudinal strength constraint is composed of a shear stress constraint and a bending moment constraint. These constraints are expressed by the equations in Figure 17. If these constraint conditions are not satisfied, a penalty is added to the evaluation function H.

[0053] The shear stress SS(n) and bending moment SB(n) are generated at frame position FP(n) by the lightship weight and buoyancy under specified draft, trim, and still water. These values ​​can be calculated from the DM (Draft Mean) and trim (TRIM[M]) of a ship loaded with a weight close to the displacement at the specified draft (DRAFT[M]) and the coefficients (SS-Coeff, SB-Coeff) at each frame position. Therefore, SS(n) and SB(n) are expressed by the calculation formula in Figure 18 described later.

[0054] Figure 18 is a diagram explaining the procedure for calculating the shear stress SS(n) and bending moment SB(n). In the calculation formula in Figure 18, TRIM is a value that is constrained by the posture constraint, and includes the state variable q in the process of searching for the optimal solution. DM is also calculated using TRIM, so it similarly includes the state variable q. In other words, SS(n) and SB(n) are calculated in the process of searching for the optimal solution.

[0055] The coefficient DT includes DM**2, and the calculation formula for DM×TRIM includes TRIM**2. The calculation formula for TRIM includes the state variable q. Therefore, the calculation of DM**2 and DM×TRIM in the constraint formulas for shear stress and bending moment includes the square of the state variable q. As a result, a large amount of memory resources is required when expanding the longitudinal strength constraint formulas in the process of searching for the optimal solution, and the longitudinal strength constraints, which require 40 constraint formulas in particular, are not practical in terms of computer resources, execution time, and communication time. Therefore, in this embodiment, the order of the longitudinal strength constraint formulas is reduced from quadratic to linear by approximating TRIM**2 to a linear formula (Maclaurin expansion).

[0056] Specifically, on the premise that TRIM is directed within a specified range, the TRIM range is set to the range shown in FIG. 18, and TRIM**2, DM**2, and DM×TRIM are approximated by the linear expression for TRIM shown in FIG. 18.

[0057] Fig. 19 is a process flow diagram explaining the process of searching for an optimal solution of the state variable q. Since the loading plan in this embodiment is also a knapsack problem, when the maximum draft is specified as a constraint, the draft moves to the maximum draft so that the cargo amount is maximized. For example, when loading raw materials with a large loading coefficient, if the hold becomes full before the maximum draft is reached, ballast water to make up for the shortage is injected. In this case, surface layer sorting, which will be described later, may be performed.

[0058] FIG. 20 is a schematic diagram for explaining surface layer sorting. In the process of repetition in FIG. 19, while some state variables are determined, the grain size of the baggage slices for undetermined state variables may be gradually reduced and repetition may be performed. For example, the baggage stowage below the surface layer may be fixed as one slice per hold, and the baggage stowage below the surface layer (each 1000m) may be fixed as one slice per hold. 3 about) into small grain slices (e.g. 500m 3 ) and conduct the search again. The amount of ballast water injected is evaluated to be decreasing and is transferred to the baggage hold, so the amount of injection at that time is set as the maximum and the slice is refined (for example, 500m 3 This solves the problem of arranging finely grained surface slices in the margins of each hold, improving accuracy while keeping or reducing the number of variables.

[0059] <First embodiment: Summary> In the stowage method according to this embodiment, holds, ballast water tanks, cargo, and ballast water are each divided into two types, large slices and small slices, and an optimal solution is searched for while satisfying maximum and minimum quantity constraints. This makes it possible to solve the stowage plan as an optimization problem within a realistically operable time while suppressing consumption of memory resources.

[0060] The loading method according to this embodiment is based on the loading evaluation coefficient CL ij ×AMP LOADING and loading / unloading evaluation coefficient CD ij ×AMP DISCHARGE Using this, we evaluate whether the loading and unloading sequences are efficient. This makes it possible to create a loading plan that takes into account the efficiency of each loading and unloading operation.

[0061] The stowage method according to this embodiment is to load cargo of the same type as the cargo loaded on the bow side or the lower level with a cargo type evaluation coefficient CT ijThis allows optimization that takes into account the efficiency of loading and unloading work based on the loading order of cargo types. The layout relationship between holds is also evaluated using the hold slice evaluation coefficient A nk can be similarly evaluated using

[0062] The loading method according to this embodiment calculates the allowable maximum weight and the required minimum weight using an approximation formula that does not include the state variable q, and calculates the penalty function H A This makes it possible to obtain a realistic stowage plan that takes hold weight constraints into account without reducing the accuracy of the stowage plan.

[0063] The loading method according to the present embodiment uses the draft constraint and the trim constraint to calculate the penalty function H A By determining the loading conditions that satisfy these constraints, assuming a specified draft, a realistic loading plan that simultaneously satisfies other constraints can be obtained.

[0064] In the stowage method according to the present embodiment, the penalty function H A A longitudinal strength constraint is used as part of the equation. In the longitudinal strength constraint, the shear stress SS(n) and bending moment SB(n) are calculated by an approximation using a linear expression for the state variable q. This makes it possible to keep the consumption of computational resources within a realistic range and obtain a planning method that is practical for actual operation.

[0065] <Embodiment 2> 21 is a configuration diagram of a stowage apparatus 1 according to the second embodiment of the present invention. The stowage apparatus 1 is an apparatus that implements the stowage method described in the first embodiment. The stowage apparatus 1 includes a calculation unit 11 and a storage unit 12. The calculation unit 11 can also receive instructions for the stowage apparatus 1 (e.g., fixed designation in fixed slice constraints).

[0066] The calculation unit 11 is a functional unit that implements the stowing method described in embodiment 1. The calculation unit 11 can be configured by hardware such as a circuit device that implements the function, or can be configured by a calculation device such as a CPU (Central Processing Unit) executing software that implements the function.

[0067] The memory unit 12 is a storage device that stores data used by the calculation unit 11. For example, it can store data describing the initial values ​​of the baggage slices and hold slices, data describing the evaluation function H and its coefficients, data describing the baggage and holds specified in the fixed slice constraints, etc. It can also store the results of calculations performed by the calculation unit 11.

[0068] The calculation unit 11 outputs data describing the stowage state obtained by carrying out the stowage method by any method. For example, the calculation unit 11 can output data describing the results, present the results on an output device such as a display provided in the stowage apparatus 1, or output the data in any other manner.

[0069] <Embodiment 3> In a third embodiment of the present invention, optimization of parameters in the above-mentioned embodiments is considered. Since the stowing method itself is similar to the above-mentioned embodiments, the present embodiment mainly considers optimal parameter values.

[0070] <Third embodiment: Number of slices that can be practically used> In order to treat a stowage plan as a knapsack problem, the cargo to be loaded must be divided into a certain amount of grains (slices) to create a finite set. Naturally, the smaller the slice width (volume or weight) is, the smaller the error will be, and the higher the accuracy of the maximum amount to be loaded will be. However, the number of slices ((cargo slices + ballast water slices) x (hold slices + ballast tank slices)) is the number of variables in the simulated annealing calculation, and it has a significant effect on the memory consumption when generating logical model objects, the transfer time of those objects, the simulated annealing calculation time, and so on. For example, for a ship with a displacement of 200,000 tons, a slice width of 1,000m would be 3 If simulated annealing is performed with a scale of about 0.5% of the displacement, the number of variables will be at least 200 x 200 = 40,000, making practical operation difficult with currently available computational resources. Therefore, in the above embodiment, the slices are divided into two types: large slices and small slices.

[0071] For example, a small slice of 1000m 3 (0.5% of the displacement) 3 The policy for dividing the slices is to divide 2 / 3 of the hold into large slices and the remaining 1 / 3 into small slices. For example, 24,000m 3 If the hold is the maximum capacity hold, divide it into 10 slices as follows. This simply divides the 1000m 3 The number of slices can be reduced from 24 to 10 when the number of variables is increased by 2.4 times. If the number of variables is increased by 2.4 times, the memory consumption will be more than 30 times (2.4 to the power of 4). Therefore, by reducing the number of slices, the memory consumption can be significantly reduced.

[0072] 16,000m 3 is 8000m 3 2 large slices The remaining 8000m 3 is 1000m 3 8 small slices of

[0073] The number of slices (large slices, small slices) is determined based on the number of holds, hold capacity, maximum allowable weight, and cargo volume for each ship, but since there is a practical upper limit to the number of variables, an appropriate number must be determined taking into consideration the number of constraints (load). For example, for a 200,000-ton ship with 9 holds, the following is considered appropriate based on memory consumption and transfer time for model objects. In this case, the number of variables is 129 x 58 = 7482 variables.

[0074] Hold slice count (2,7) Ballast tank slice count (1,7)

[0075] The size of the slice is determined by the following policy. This specifies the division size in terms of capacity, but the division size in terms of weight is also determined based on the maximum allowable weight within the same division number (2,7).

[0076] ·num_of_small_slice×size_of_small_slice≧size_of_big_slice ·Σ(small_slices+big_slices)≧max(hold capacity)

[0077] <Embodiment 3: Capacity and weight> In the above embodiment, by providing two attributes, capacity and weight, to the baggage slices, it is possible to apply hold capacity constraints to baggage that is volume-oriented (described later), and to apply weight constraints to baggage that is weight-oriented (described later). Ballast water is treated as baggage that is volume-oriented. In other words, the division size of the baggage and ballast water changes depending on whether their SF (stowage factor) is larger or smaller than the boundary SF value. A plurality of types of baggage slices with different capacities and weights are mixed in the holds and ballast tanks that are divided into large slices and small slices.

[0078] Cargo with a high SF, such as coal, is divided into sizes determined from the maximum hold capacity, and the weight equivalent to that capacity is assigned as an attribute. Cargo with a low SF, such as iron ore, is divided into sizes determined from the maximum hold weight, and the capacity equivalent to that weight is assigned as an attribute.

[0079] "Cargo that wins by weight" is cargo that exceeds the maximum allowable weight of a hold before the hold becomes full. "Cargo that wins by volume" is cargo that fills a hold before exceeding the maximum allowable weight of the hold. Therefore, the SF that is the boundary between the two judgments differs for each ship, but for example, the SF that is the boundary can be calculated as follows:

[0080] The boundary SF[CF / LT]=(max(hold capacity[m 3 ])x40x1.016) / (max(Maximum allowable weight of hold [MT])x1.1327)

[0081] The most important point of this method is as follows. The holds and ballast tanks are 129 squares consisting of 9 holds (81 squares) divided by the number of slices (large, small) = (2, 7) and 8 ballast tanks (64 squares) divided by the number of slices (large, small) = (1, 7), and the squares themselves do not have capacity or weight. The only rule that slices configured in this way must satisfy is that "large baggage slices can only enter large hold slices, and small baggage slices can only enter small hold slices." This makes it possible to apply both capacity and weight constraints at the same time, and it is possible to handle cargo that is weight-first, cargo that is volume-first, or a mixture of both, with the same model (formulation).

[0082] <Third embodiment: Regarding errors caused by slice division> In the above embodiment, for example, the displacement Δ=167786 [MT] at the specified draft DCF=15.00 [M] is read from the Hydrostatic Table, and under the constraint of "loading the largest cargo whose dead weight (DW: deadweight) is closest to Δ", the trim is calculated using the LCB (Longitudinal Center of Buoyancy) and MTC (moment required for 1 cm trim) also read from the Hydrostatic Table at DCF=15.00 [M], and a trim constraint is imposed so that the trim falls within the specified trim range.

[0083] This procedure generates errors in trim calculations and subsequent strength calculations when DW has not sufficiently converged to Δ. Therefore, we decided to repeatedly perform annealing while always maintaining sufficient accuracy (e.g., DW ≥ 0.999Δ, with ballast water below a specified amount as a guideline).

[0084] This error is caused by the fact that the cargo, ballast water, holds, and ballast water tanks are sliced ​​into small volumes to some extent and made into finite sets in order to solve the loading combination problem as the "multiple knapsack allocation problem." The error can be further reduced by implementing the surface sorting described in the first embodiment.

[0085] <Modifications of the present invention> In the above embodiment, simulated annealing has been exemplified as an algorithm for solving the stowage planning problem, but any other optimization algorithm may be used. The stowage apparatus 1 may be configured as, for example, an Ising machine. [Explanation of symbols]

[0086] 1: Stacking equipment 11: Arithmetic section 12: Storage part

Claims

1. 1. A method for determining a sequence for loading cargo onto a ship carrying cargo and ballast water, comprising the steps of: The ship includes a hold that is a container for carrying the cargo, and a ballast tank that carries the ballast water, The stacking method comprises: A step of virtually dividing the hold into two types of hold slices, a large hold slice and a small hold slice having a capacity smaller than that of the large hold slice, in a height direction of the ship; A step of virtually dividing the ballast tank into two types in the height direction: a large ballast tank slice and a small ballast tank slice having a capacity smaller than that of the large ballast tank slice; defining a function for calculating an evaluation value of the loading state for a loading state in which the cargo is loaded in at least one of the large hold slice or the small hold slice and the ballast water is loaded in at least one of the large ballast tank slice or the small ballast tank slice; calculating the evaluation value according to the function; searching for the mounting state that can optimize the evaluation value; having The function is a maximum quantity constraint that imposes a penalty that worsens the evaluation value if the quantity of the cargo to be loaded onto the ship exceeds a maximum contract quantity; a minimum quantity constraint that imposes a penalty that worsens the evaluation value if the quantity of the cargo loaded onto the ship is less than a minimum contract quantity; It is made up of A loading method characterized by the above.

2. the minimum quantity constraint specifies, for each type of luggage, a minimum quantity of the luggage to be loaded onto the large hold slice and a minimum quantity of the luggage to be loaded onto the small hold slice; In the step of dividing the hold, the hold is divided so that the number of the large hold slices is equal to or greater than the total number of the minimum quantities of baggage to be loaded onto the large hold slices for all baggage types, and the number of the small hold slices is equal to or greater than the total number of the minimum quantities of baggage to be loaded onto the small hold slices for all baggage types.

2. The method of claim 1, wherein the loading is performed in a manner similar to that described in claim 1.

3. the minimum quantity constraint specifies a minimum quantity of the ballast water to be loaded into the large ballast tank slice and a minimum quantity of the ballast water to be loaded into the small ballast tank slice; In the step of dividing the ballast tank, the ballast tank is divided so that the number of the large ballast tank slices is equal to or greater than the minimum quantity of ballast water to be loaded onto the large ballast tank slices, and the number of the small ballast tank slices is equal to or greater than the minimum quantity of ballast water to be loaded onto the small ballast tank slices.

2. The method of claim 1, wherein the loading is performed in a manner similar to that described in claim 1.

4. In the step of defining the function, the function is defined for the loading state in which a first baggage is loaded onto a first hold slice generated in the step of dividing the hold, and a second baggage is loaded onto a second hold slice located below or toward the bow of the first hold slice; The function is a luggage slice evaluation factor evaluating a relationship between the first luggage in the loaded state and the second luggage in the loaded state; a hold slice evaluation coefficient for evaluating a relationship between the first hold slice in the mounted state and the second hold slice in the mounted state; It is made up of 2. The method of claim 1, wherein the loading is performed in a manner similar to that described in claim 1.

5. The function is a binary value representing whether the first load is loaded on the first hold slice; a binary value representing whether the second load is loaded on the second hold slice; The baggage slice evaluation coefficient, the hold slice evaluation coefficient, and configured to calculate the evaluation value by multiplying 5. The method of claim 4, wherein the loading is performed in a manner similar to that described in claim 4.

6. The baggage slice evaluation coefficient is composed of a loading evaluation coefficient that evaluates the loading order of the first baggage and the second baggage, The loading evaluation coefficient is configured to have an effect of improving the evaluation value when a loading port of the second baggage is prior to a loading port of the first baggage, The loading evaluation coefficient is configured to have an effect of worsening the evaluation value when a loading port of the second baggage is later than a loading port of the first baggage, The loading evaluation coefficient is configured to have no effect of improving or worsening the evaluation value when the loading port of the second baggage is the same as the loading port of the first baggage.

5. The method of claim 4, wherein the loading is performed in a manner similar to that described in claim 4.

7. The baggage slice evaluation coefficient is composed of a loading / unloading evaluation coefficient for evaluating a loading / unloading order of the first baggage and the second baggage, The loading / unloading evaluation coefficient is configured to have an effect of improving the evaluation value when the loading / unloading port of the second baggage is later than the loading / unloading port of the first baggage, The loading / unloading evaluation coefficient is configured to have an effect of worsening the evaluation value when the loading / unloading port of the second baggage is prior to the loading / unloading port of the first baggage, The loading and unloading evaluation coefficient is configured to have no effect of improving or worsening the evaluation value when the loading and unloading port of the second baggage is the same as the loading and unloading port of the first baggage.

5. The method of claim 4, wherein the loading is performed in a manner similar to that described in claim 4.

8. The baggage slice evaluation coefficient is composed of a baggage type evaluation coefficient for evaluating the first baggage type and the second baggage type, The baggage type evaluation coefficient is configured to have an effect of improving the evaluation value when the type of the first baggage and the type of the second baggage are the same, The baggage type evaluation coefficient is configured to have an effect of worsening the evaluation value when the type of the first baggage is different from the type of the second baggage, The baggage type evaluation coefficient is configured so as to have no effect of improving or worsening the evaluation value for the ballast water.

5. The method of claim 4, wherein the loading is performed in a manner similar to that described in claim 4.

9. The vessel includes a plurality of holds arranged along a fore-aft direction of the vessel, The hold slice evaluation coefficient is configured to have an effect of worsening the evaluation value by a first evaluation amount when loading into the hold located on the stern side of the ship within the hold slice evaluation coefficient occurs before loading into the hold located on the bow side of the ship within the hold slice evaluation coefficient.

5. The method of claim 4, wherein the loading is performed in a manner similar to that described in claim 4.

10. The hold slice evaluation coefficient is configured to have an effect of worsening the evaluation value by a second evaluation amount that worsens the evaluation value more significantly than the first evaluation amount when different types of baggage are mixed in the same hold.

10. The method of claim 9,

11. the function is configured by a penalty function that worsens the evaluation value when the order does not satisfy a constraint condition, The penalty function is configured to impose a penalty that worsens the evaluation value when the total volume of the luggage loaded in the hold exceeds the loadable capacity of the hold.

2. The method of claim 1, wherein the loading is performed in a manner similar to that described in claim 1.

12. the function is configured by a penalty function that worsens the evaluation value when the order does not satisfy a constraint condition, The penalty function is a weight constraint of the baggage carried by the hold, a single hold maximum weight constraint defining the maximum amount of cargo that a single hold can carry; a single hold minimum weight constraint defining the minimum amount of cargo that a single hold must carry; an adjacent hold maximum weight constraint that defines the maximum combined weight of the cargo that two adjacent holds can carry; an adjacent hold minimum weight constraint that specifies the minimum total weight of the cargo that two adjacent holds must carry; It is made up of 2. The method of claim 1, wherein the loading is performed in a manner similar to that described in claim 1.

13. The method further comprises: calculating the draft of the hold using a draft corresponding to the displacement of the vessel in the loaded state, a maximum trim of the vessel, a longitudinal center of buoyancy of the vessel, and a longitudinal center of gravity of the hold; calculating the weight constraint imposed on the hold using the draft of the hold; having the step of calculating the weight constraint imposed on the hold includes calculating the allowable load weight in the hold by fitting the draft of the hold to a linear relationship between the allowable maximum load weight of the vessel and the allowable minimum load weight of the vessel; The step of calculating the weight constraint imposed on the hold includes calculating the required cargo weight in the hold by fitting the draft of the hold to a linear relationship between the required maximum cargo weight of the vessel and the required minimum cargo weight of the vessel.

13. The method of claim 12.

14. The penalty function is configured such that the weight of the adjacent hold maximum weight constraint and the weight of the adjacent hold minimum weight constraint are half the weights of the other constraints.

13. The method of claim 12.

15. the function is configured by a penalty function that worsens the evaluation value when the order does not satisfy a constraint condition, The penalty function is configured such that a solution is obtained in the searching step such that a difference between a total weight of the vessel in the loaded state and a displacement of the vessel in the loaded state is less than or equal to a threshold value.

2. The method of claim 1, wherein the loading is performed in a manner similar to that described in claim 1.

16. the function is configured by a penalty function that worsens the evaluation value when the order does not satisfy a constraint condition, The penalty function is configured such that the searching step results in a solution that results in a trim of the vessel at the on-board condition within a specified range.

2. The method of claim 1, wherein the loading is performed in a manner similar to that described in claim 1.

17. The stowage method further comprises the step of calculating shear stress and bending moment for each longitudinal position of the ship; the function is configured by a penalty function that worsens the evaluation value when the order does not satisfy a constraint condition, The penalty function is comprised of a longitudinal strength constraint that requires that the shear stresses at all the locations and the bending moments at all the locations be within an allowable range.

2. The method of claim 1, wherein the loading is performed in a manner similar to that described in claim 1.

18. The stowage method further comprises the step of calculating a trim of the vessel in the loaded condition; In the step of calculating the shear stress and the bending moment, the trim calculated in the step of calculating the trim is used to calculate the shear stress and the bending moment; In the step of calculating the shear stress and the bending moment, the trim squared term in the calculation formula is approximated by the trim first power term, and the shear stress and the bending moment are calculated using the approximate formula.

18. The method of claim 17.

19. The method further comprises: a step of redefining the remaining slices, excluding the large hold slice or the small hold slice located at the top of the holds, as a single slice for each of the holds, when any slices are left behind as a result of the searching step; reducing the leftovers by dividing the large hold slice or the small hold slice located at the top of the hold into smaller slices and re-running the search; have 2. The method of claim 1, wherein the loading is performed in a manner similar to that described in claim 1.

20. 1. A loading system for determining a sequence for loading cargo onto a ship carrying cargo and ballast water, comprising: The ship includes a hold that is a container for carrying the cargo, and a ballast tank that carries the ballast water, The stowage device includes a calculation device that determines the order, The computing device includes: A step of virtually dividing the hold into two types of hold slices in the height direction of the ship, the large hold slices and the small hold slices having a capacity smaller than that of the large hold slices; A step of virtually dividing the ballast tank into two types in the height direction: a large ballast tank slice and a small ballast tank slice having a capacity smaller than that of the large ballast tank slice; defining a function for calculating an evaluation value of the loading state for a loading state in which the cargo is loaded in at least one of the large hold slice or the small hold slice and the ballast water is loaded in at least one of the large ballast tank slice or the small ballast tank slice; calculating the evaluation value according to the function; searching for the mounting state that can optimize the evaluation value; determining said order by carrying out The function is a maximum quantity constraint that imposes a penalty that worsens the evaluation value if the quantity of the cargo to be loaded onto the ship exceeds a maximum contract quantity; a minimum quantity constraint that imposes a penalty that worsens the evaluation value if the quantity of the cargo loaded onto the ship is less than a minimum contract quantity; It is made up of A stacking device characterized by:

21. A stowage program that causes a computer to execute the stowage method according to any one of claims 1 to 19.

Images