KR20090115274A - Method for planning load sequence of containers - Google Patents

Abstract

PURPOSE: A method for planning load sequence of containers is provided to increase the workability of shipping work by introducing various estimation elements for the container shipping work. CONSTITUTION: Workable slots are selected in a loading target slot group of a ship bay(S202). Candidate containers of searched each slot are selected(S204). By the combination of selected slots and candidate containers, candidate nodes are generated(S206). The generated all of the candidate nodes are evaluated(S207). Based on the estimation, nodes are selected as much as a beam width(S208). The evaluation is calculated according to each node. One node is selected among the generated beam nodes.

Description

{Method for planning load sequence of containers}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to container ship operations, and more particularly, to an individual container loading order planning method capable of increasing the productivity of the ship operations by introducing various evaluation elements to establish individual container loading order planning.

Recently, with the increase in the volume of transportation between countries, shipping companies are demanding the rapid treatment of the increased volume.

Therefore, the container terminal is trying to reduce the logistics cost by reducing the productivity of loading and unloading work for ships, and shorten the ship's flight time. As a result, the work order for each efficient container must be planned.

Container terminals are divided into a quay wall, a storage area and a gate area.

At the container terminal, the unloading operation of loading and unloading the import container loaded on the ship, the loading operation of loading the export container loaded on the loading dock, and the external container gated the import container loaded on the loading dock. It consists of an export operation withdrawal through and an import operation with receiving export containers from an external truck through a gate and loading them into the storage.

Among these, unloading and unloading are often referred to as ship operations. In order to carry out this onboarding work, planning is necessary.

In general, before the vessel is docked at the berth, the planner schedules the work of the quay-walled crane (QC) based on a shipping plan received from the shipping company and a yard map showing the status of the container at the dock. Establish. In addition, device positioning for individual unloading containers and the withdrawal order of the individual unloading containers are determined.

The plan for unloading is a process of finding and withdrawing individual containers from the yard and loading them in a designated position on the ship, which requires more consideration and planning time than planning for unloading operations that only require space.

As container traffic volume increases, it is necessary to establish a container loading order plan that can increase work efficiency by considering the situation of the plant floor and the ship's working situation at the time of loading work.

However, at present, there is no method to plan the loading sequence considering the situation of various sites during container loading operation.

The container loading order plan of the prior art is established in consideration of any one of the vessel work efficiency or the plant work efficiency, or is established in consideration of the ship work efficiency and the plant work efficiency in order to increase the work productivity.

In addition, the operation planning system of the container terminal of the prior art has a problem of inefficient because the planner's intervention is required at the time of dripping work planning.

The present invention is to solve the inefficient problem that requires a lot of planner's involvement in the establishment of the loading operation plan to apply to the operation plan system of the container terminal of the prior art, the individual container loading sequence by introducing various evaluation elements The objective is to provide a method for planning individual container loading orders that will increase the productivity of the ship operations by planning.

An object of the present invention is to provide an individual container loading order planning method that can increase productivity by considering the main operation and the yard work at the same time when establishing the individual container loading order plan affecting the productivity of the container terminal. .

The present invention aims to increase productivity by introducing various evaluation factors to establish individual container loading order plans and using them in an operation planning system of container terminals.

The individual container loading order planning method according to the present invention for achieving the above object, in order to establish the container loading order planning, create the initial beam node, search all the beam nodes, work in the loading target slot set of the ship bay Performing a process of finding the available slots and finding a candidate container of the found slots, and generating candidate nodes with a combination of the found slots and respective elements of the candidate containers; evaluating all generated candidate nodes and Selecting a node by a beam constant based on the evaluation calculated for each node; selecting one node among the beam nodes generated by the selection, and dropping from the initial node by backtracking the node as an end point Determining a work order.

Such individual container loading order planning method according to the present invention has the following effects.

In the individual container loading order planning method according to the present invention, by introducing various evaluation factors for the container final operation, the productivity of the main operation can be improved by establishing the individual container loading order planning.

In other words, by providing algorithms reflecting on-site constraints and evaluations that are considered in actual container terminals, it is possible to automate the loading operation.

Hereinafter, a preferred embodiment of the individual container loading order planning method in consideration of various evaluation factors according to the present invention will be described in detail.

Features and advantages of the individual container loading order planning method according to the present invention will become apparent from the following detailed description of each embodiment.

1A to 1C are diagrams showing an example of a quay-wall crane work schedule plan, and FIG. 2 is a diagram showing a cross-sectional view of a ship as an example of a shipping instruction, and FIG. 3 is a yard map. Fig. 1 shows an example of the configuration.

The present invention determines the selection and shipping order of individual containers for loading operations using the work schedule of a given quay-walled crane, and this order planning is referred to as individual container loading order planning.

At this time, the yard crane (YC) is in charge of pulling out the container from the installation site, and the quay-wall crane (QC) is in charge of placing the container at a specific position of the vessel.

The loading operation of the ship operation is the operation of loading a container which is required by the ship on a container already loaded in the yard. This should be considered at the same time as the site situation and the ship operations. The present invention summarizes the above two requirements into four main variables and proposes an algorithm for determining the loading sequence which considers them simultaneously.

(Information used for individual container loading plans)

The individual container loading plan according to the present invention is to determine the working order and shipping position of individual containers to be loaded on a ship. In this individual container loading plan, three pieces of information are used:

In other words, the individual container loading plan is established by using QC work crane work schedule, stowage plan, and yard map information.

First, the QC work schedule will be described.

The basic unit of QC work is a ship cluster, which is defined as a collection of containers that have the same attributes (destination port, size, type, quantity / load) and are gathered in the same location.

However, in order to maximize work efficiency, a single QC can simultaneously load containers with different properties if possible. Therefore, in the present invention, it is assumed that the cluster is mainly generated in units of ship bays and is divided into hold or deck.

Thus, the QC work schedule can be seen as determining which cluster is responsible for which QC and what time zone the cluster will perform.

1A is an example of a shipping instruction, and FIG. 1B shows an example of a method of dividing a cluster according to a shipping instruction.

1C shows an example of performing a QC work schedule in units of clusters. One cluster contains containers for different destination ports, and is assigned a number of fetches per block.

The description of the Stowage Plan is as follows.

The shipping instructions are sent from the shipping company to the terminal, which contains information about the individual containers to be loaded on board the ship. 2 is an example of a shipping instruction for a particular ship bay.

The yard map is described as follows.

The yard map shows container loading information in the storage area, and can be viewed as a map showing the location information and the attribute information of individual containers stored in the storage area.

In general, it is expressed in a yard bay unit and is represented as shown in FIG.

As described above, the individual container loading plan is established by using the QC Work Crane work schedule, Stowage plan, and yard map information. Consider the points.

(Considerations when establishing individual container loading plans)

Description of the container loading planning method according to the present invention described below will be described based on the following criteria.

First, only the loading work.

Second, since the size of the problem is limited to one ship cluster, only one quay-crane is considered.

Third, all containers required for the unloading operation will be described based on the state stored in the equipment storage.

It is assumed that the direct-loading container which is loaded directly without being stored in the storage space is pre-stored in a virtual dummy block.

First, considerations related to work preferences are related to the QC work pattern, container weight class, load balancing between plant blocks, and rehandling.

First, consider the QC work pattern.

QC work is oriented for operator convenience, ship safety and work efficiency. For example, when working on hold, work is done on another stack after the work is completed on the same stack. This is due to the convenience of work. When selecting the next stack, the ship's stability is taken into account. When the work is done on the deck, the work of the upper tier is performed after the lower tier work is completed. This is due to the efficiency of the work.

Second, consider the container weight class.

Container weight is an important factor when developing loading plans. Since weight affects the stability of the ship, a plan should be developed taking into account the weight distribution of the containers loaded throughout the ship. In general, shipping instructions from the shipping company describe the weight.

Third, consider the load balancing among the device length blocks.

Consideration should be given to space and handling resources within the block. By distributing work that is concentrated on a particular block, you can distribute the workload of the block and improve productivity on board.

Fourth, consider the matter of rehandling.

When another container exists on the target container, the top container needs to be moved for the pulling out of the target container. This is called rehandling. This rehandling can lead to delays in QC working time and adversely affect the onboard productivity. Therefore, loading plans should be established to minimize rehandling.

(Restrictions to be followed when establishing individual container loading plans)

Restrictions that must be observed when establishing a container loading plan according to the present invention include conformity with the shipping instructions, loading restrictions on board, and loading restrictions on board.

First, the correspondence with the shipping instructions should be considered. For each slot described in the shipping instructions, a container must be assigned that matches the information defined.

Second, consideration should be given to the loading constraints on the deck, where the ship bay is divided into onboard and onboard, and a hatch cover exists between and onboard.

On board, hatch covers cannot be loaded in excess of the weight they can bear. Therefore, the maximum weight allowed per stack on board cannot be exceeded.

Third, consideration should be given to the loading constraints on hold, which may not exceed the maximum height allowed for each stack. If exceeded, the hatch cover which distinguishes the ship from the ship cannot be closed.

Based on the considerations described above, the container loading planning process will be described in detail.

4 is a flowchart illustrating a container loading order planning method in consideration of various evaluation factors according to the present invention.

The container loading order planning method according to the present invention is established using a beam search technique.

The beam search technique is a variation of breadth first search, and the solution search is performed while keeping the width of the beam, which is the search space, constant.

In other words, a heuristic search technique that constructs a decision tree, extends the branches to only some nodes that have the potential to do good harm at each level, and eliminates the remaining nodes, thereby reducing the computation time and reducing computation time. to be.

The number of nodes in each step is determined by the beam width. Therefore, it is important to determine the appropriate evaluation function and beam constant for the problem.

Each node is created from a combination of a ship's slot and a container candidate of the device head that can be installed in that slot. In each step, the beam nodes determine the best (lowest) order of evaluation by the number of beam constants.

At this time, at least one child node is selected at each beam node in order to prevent falling into the local solution.

The final solution selects the node with the best (smallest) cumulative evaluation value for the beam nodes determined in the last step. Backtracking the node as an end point determines the loading sequence from the initial node.

Definition of each element used in the container loading order planning method in consideration of various evaluation factors according to the present invention is as follows.

sign meaning Ψ Set of job candidate slots Φ The set of containers that can work in the storage block for one slot of Ψ Ω Set of nodes (created by a combination of slots and field-length containers) created using each element of Ψ and Φ d Beam constant r Level at which the beam node is determined z Total number of slots in the set of loaded slots

First, an initial beam node is generated (S101), and processing of slot collection, slot selection, and candidate container selection that can be performed for each beam node is performed (S102).

Specifically, the beam node processing process will be described.

When the beam node is generated, the beam node that has not been processed above is selected among the current beam nodes (S201).

And collects a slot available for operation (S202).

The collection of workable slots initializes the workable container set (Φ) in the device-length block for one slot of the work candidate slot (Ψ) and the set of work candidate slots (Ψ).

It finds all available slots in the current loading target slot set. In other words, the lowest level slot is searched at the current level. Add all selected slots to (Ψ).

Next, one slot is selected from all slots in (Ψ). (S203)

Then, a candidate container of a device length is selected for the selected slot (S204).

Here, the candidate container selection of the device length finds a container having the same attribute as the corresponding slot at the top of each block of the device length. Only the topmost containers should be covered to minimize rehandling.

If it can't find a container, it will rework and find a container that meets the conditions. Add all containers that meet the condition to (Φ) and delete the corresponding slot in (Ψ).

This container selection process is repeated until all slots have been made (S205).

The node is created through the combination (S206), that is, the node is created through the combination of the elements of (Ψ) and (Φ). The generated nodes are added to the set of nodes (Ω) generated through the combination of the elements of (Ψ) and (Φ).

The above-described processing is repeated until processing for all beam nodes is made (S103).

Subsequently, when processing is performed on all beam nodes, evaluation of all generated nodes is performed (S207).

Here, the evaluation method of each generated node is applied to the penalty by assessing the items of the matching of the QC work pattern, the difference in the weight class of the slot and the container to be loaded, the load smoothing between blocks, the rehandling work in the equipment field, and the penalty sum. Determined by the value.

When all nodes generated in this manner are evaluated, nodes are selected by the beam constant (S208).

That is, the best nodes are selected based on the evaluation values calculated for each node.

When selecting a node, at least one child node is selected at each beam node in order to avoid falling into local solutions. The selected d nodes become beam nodes. It is determined whether r ≥ z to determine whether all the loading target slots have been processed (S209).

If r ≥ z, perform the step of selecting the final solution (S210), otherwise increase r by 1, initialize (Ω), and perform step S102.

In the final r-th step, d beam nodes are generated. You must select one of these nodes. The final beam node selects the node with the best cumulative evaluation value. Backtracking the node as an end point determines the loading sequence from the initial node.

The evaluation method of each of the nodes generated in the above step S207 will be described in detail.

First, a method of scoring the work pattern of the quay wall crane and reflecting it in the evaluation item is as follows.

The work pattern of the quay-crane defines the loading direction of the container when the quay-crane is loading.

Work patterns are classified into four work patterns as shown in FIG. 5.

The work on board is a stack-level work, so there are three ways to impose a penalty. The sea side is divided into the land side or the land side in the sea direction work and the work in the direction of spreading from the middle to both sides.

The penalty calculation method is a rank given to each stack including a candidate slot divided by the number of target stacks.

For example, if the total number of stacks is 6 and the pattern of working from sea side to land side is applied, the penalty is applied in the order of 1/6, 2/6, etc. from the first sea side stack. Work is advantageously done in small order, so scoring and reflecting in this way can easily be reflected in the evaluation.

Because work on the deck works on a tier basis, it only penalizes the difference between the tiers.

In addition, the method of scoring the weight grade difference between the slot and the container and reflecting it in the evaluation item reflects the planner by dividing the weight grade as shown in FIG. 6A and the penalty value according to the weight grade difference as shown in FIG. 6B.

And the method of imposing a penalty on load leveling among blocks and reflecting it in the evaluation items is as follows.

The load leveling object is limited to blocks corresponding to the ship cluster. Each block stores a different number of containers for loading into the cluster.

If the load is carried out continuously in a particular block, the workload of that block will increase. To prevent this, the average workload of each block is calculated and the task is planned by selecting a block with a small difference between the cumulative load and the reference load. Use this difference as a penalty.

Here, the equation for obtaining the reference load is defined as in FIG.

는 k회차 적하작업시 블록i에서의 누적 적하 작업 컨테이너 수

,

는 QC 작업 일정 계획에서 블록 i에 할당된 적하 컨테이너 개수로 정의되고, 블록 i의 기준 부하는

이고, k회차 적하 작업시 블록 I에서의 부하 평활화에 대한 패널티는

으로 정의된다.n is the total number of slots in the cluster, k is the number of loading operations,

Is the cumulative number of stacked job containers at block i for k times of load.

,

Is defined as the number of load containers assigned to block i in the QC work scheduling, and the reference load for block i

The penalty for load smoothing at block I in k times

Is defined.

And the method of imposing penalties for rehandling at the equipment site and reflecting them in the evaluation items is as follows.

이 된다.The penalty value is determined by the expected number of rehandling. In other words, if the total number of stackable stages is k,

Becomes

This is determined by dividing the expected number of rehandling times by the target container. 8 summarizes the penalty according to the expected number of rehandling.

The container loading order planning method according to the present invention can increase the productivity of the ship operation by introducing a variety of evaluation factors for the container ship operation to establish the individual container loading order planning.

Those skilled in the art will appreciate that various changes and modifications can be made without departing from the spirit of the present invention.

Therefore, the technical scope of the present invention should not be limited to the contents described in the embodiments, but should be defined by the claims.

Figures 1a to 1c is a configuration diagram showing an example of the quarry wall work schedule plan

2 is a cross-sectional view of the ship as an example of a shipping instruction

3 is a block diagram showing an example of a yard map (Yard map)

4 is a flow chart for a container loading order planning method in consideration of various evaluation factors in accordance with the present invention

5 is a configuration diagram showing the work pattern of the quay wall crane by type

6A and 6B are tables listing penalties due to the difference between the container weight group and the weight group.

7 is a formula table for considering load smoothing between blocks when establishing a container loading order planning method according to the present invention

8 is a table depicting penalties for rehandling in a yard bay.

Claims (11)

To plan the container loading order, Create an initial beam node, search all beam nodes, Performing a process of finding slots capable of working in the loading target slot set of the ship bay and finding a candidate container of the found slots, and generating candidate nodes using a combination of the found slots and respective elements of the candidate containers; Evaluating all the generated candidate nodes and selecting nodes by a beam constant based on the evaluation calculated for each node; Selecting one of the beam nodes generated by the selection, and backtracking the corresponding node as an end point to determine a loading operation order from an initial node. The method of claim 1, wherein the search for the beam nodes comprises: Constructing a decision tree while keeping the width of the beam, the search space constant, to extend branches only to some nodes that have the potential to do good at each level. Individual container loading order planning method using a heuristic search technique to reduce the computation time by eliminating the remaining nodes. The method of claim 1, wherein in order to establish a container loading order plan, Individual container loading order planning method using QC work scheduling, shipping instructions, yard map information. The method of claim 1, wherein the process of finding a candidate container comprises: An individual container loading order planning method, which finds a container having the same property as a corresponding slot at the top of each block to minimize rehandling. The method according to claim 1, wherein in establishing a container loading order plan, Correspondence with the shipping instructions, which assigns a container that matches the information defined for each slot described in the shipping instructions; Loading constraints not to exceed the maximum weight allowed for each stack on board; A method for planning individual container loading orders, which considers loading constraints on board ships that do not exceed the maximum height allowed for each stack. The method of claim 1, wherein in evaluating all generated candidate nodes, When carrying out the loading operation, the work pattern of the quay-crane which defines the loading direction of the container is classified, and the work from the sea side to the land side, the work from the sea side to the sea side, and from the middle to both sides Individual container loading order planning method characterized by imposing a penalty by dividing the work in the direction of spreading to the evaluation items. According to claim 6, Inner work in which the penalty calculation method is performed in units of stacks, the rank given to each stack including the job candidate slot divided by the number of target stacks, A method of planning an individual container loading order, in which a work on a deck that works in tiers imposes penalties only on the difference of tiers. The method of claim 1, wherein in evaluating all generated candidate nodes, A method for planning individual container loading orders, characterized in that the planner applies scores based on criteria for dividing the weight classes and penalties according to the difference between the weight grades of the slots and containers. The method of claim 1, wherein in evaluating all generated candidate nodes, Calculate the average workload of each block that stores different numbers of containers, and select a block with a small difference between the cumulative load and the reference load to impose a penalty on the difference between the cumulative load and the reference load. Individual container loading order planning method, characterized in that reflecting. The method of claim 9, wherein the reference load and the penalty, n is the total number of slots in the cluster, k is the number of loading operations,

Is the cumulative number of stacked job containers at block i for k times of load.

,

Is defined as the number of load containers assigned to block i in the QC work scheduling, and the reference load for block i

The penalty for load smoothing at block I in k times

Individual container loading order planning method, characterized in that defined as. The method of claim 1, wherein in evaluating all generated candidate nodes, Penalties for re-handling at the site will be imposed on the assessment items, The penalty value is the total number of stackable stages.

And determining this by the value divided by the expected rehandling frequency of the stage in which the target container is located.

Images