MXPA99010866A - Method and system for automatic marker making - Google Patents

Abstract

This invention is specifically concerned with the marker making problem which arises in garment industry, i.e. the problem of placing a given number of pieces on a cloth observing certain constraints like minimum wastage. The method of automatically making markers according to this invention is characterized by the use of a goal net which consists of goals and interactors, connected to each other, and for which a first and a second direction of information flow is defined from a top level goal to a bottom level goal/an interactor and from a bottom level goal/an interactor to a top level goal, respectively.

Description

ARCATORS system for example In the industry there are "many problems that involve the problem of packaging objects in containers, in such a way that minimizes and maximizes some amount subjected to certain efforts." An example could be the packaging of products in containers to maximize the weight in the container. container or circuits packaged in a microcircuit, to maximize the computational power in the circuit while the circuit is being recalled This invention relates specifically to the manufacturing problem or marker that arises in the textile industry, i.e. , the problem of placing a given number of pieces in a fabric or saving certain efforts as minimum wear.

The problem of marking refers to the packing of two-dimensional coligones in a given area without any overlap between the polygons, so that mr. ipza the area not with enida in the collection of REF .: 32082 polygons. An interesting formulation of this problem is to pack the polygons in a given rectangle so as to maximize the number of polygons that fit in the rectangle.

Originally, the markers were created manually through the experience of bookmark makers. After the marker manufacturing process was partially automated and many attempts have been made to completely automate the process of making the marker.

For example, EP 0 664 091 Al discloses a method and system for the production of automatic markers in which the creation of a new marker is facilitated through the use of existing marker designs. A computer database of existing markers is searched for markers that are "similar" to the marker that is to be created. An existing marker is considered "similar" if it satisfies certain marker manufacturing criteria, specified by the user. Initially, the position and orientation data of the pattern pieces in the "similar" marker are used to position and orient the corresponding pieces in the new marker. Then, the new marker is "compacted" using a routine of programming elements to couple all the new pieces simultaneously, instead of coupling one or two pieces at a time. The compaction routine corrects overlaps between the parts and ensures that the same parts fit together as efficiently as possible.

It can be shown that the simplified form of the marking problem, in which the rotations and reflections of the pieces are ignored, is the complete NP. No one knows yet whether the addition of rotations increases the marking problem to a greater degree of complexity. What this statement means in practical terms is that it is almost certainly impossible to find an algorithmic method that will solve the problem of general marking at any accepted time. Therefore, an alternative for the purely algorithmic method is demanded.

In addition, the effects of a small error made at the beginning of the formation of a marker are amplified at least exponentially throughout the training process. In this way, a bad judgment can remain undetectable at the beginning of a marker for an arbitrarily long time and still make the use of a psychoanalytic impossible to achieve. In other words, there is no way to determine whether or not a developed marker will be a good marker without physically completing the marker.

Against this background, it is the object of the invention to provide an efficient system and method for the production of automatic markers.

The invention accomplishes the above objective by providing a method and system for the production of automatic markers using an objective network consisting of targets and interactors as described in the independent claims. Advantageous embodiments are described in the subclaims.

In the following method and system for making the markers automatically according to the invention, it will be explained in greater detail with reference to the drawings.Figure 1 shows an example of a marker.

Figure 2 shows part of a target network according to the invention.

- Figure 3A and 3B show jP '"examples of rectangular markers having rectangular pieces.

Figure 4 shows an ejep? Flo of the structure of the target network according to the invention.

Figures 5A to 5H show a series of markers obtained by the method and system according to the invention.

Figures 6A to 6K show an additional series of markers obtained by the method and system according to the invention.

The method and system of the invention include the concept of groupings of the pieces in the resolution of the marking problem. A marker can be understood as a simple long group of pieces that can be divided into smaller groups, for example in a clothing marker showing in Figure 1 a group of bands in the lower right and the two rows of panels falling to the right and left of the lines a. The smaller groups can be divided into even smaller groups until the individual piece is reached. % & amp; amp; amp; & amp; & amp; & ? *? £ * r. The concept of groups of leaders identifies local and global objectives. side, a group of pieces can be manipulated in. isolation without affecting other groups. Otherwise, a group of pieces can be considered within the total context of their relationship with respect to other groups. The method and system according to the invention take into account simultaneously global and local considerations implementing an objective network.

Figure 2 illustrates the structure of a target network. In general, the objective network comprises objectives and interactors. The objectives are agents that make decisions: they have no power to act, but simply delegate responsibility for the action to other agents. The interactors are operative agents: they are incapable of delegating responsibility, and instead they dedicate themselves to perform some operation on the markers themselves (for example, placement, survey, rotation of the pieces / grouping, etc.). When they have accomplished this task they pass information to the objectives so that they can initiate additional processing. A lower objective connects to the sub-objects that connect either to the additional sub-targets or to the interactors. The interactors provide the union of the target network and the marker An objective network involves two influxes of influence: motivation and satisfaction. With reference to Figure 2,. In the first sweep from left to right through the target network, the MARCADOR of the higher level objective motivates each of the BANDS and PANELS of the sub-ob ective, which in turn motivate several additional sub-objectives (p. eg DETECTION OF DRILLING, SELECTION OF PLACEMENT, etc.) or interactors (not shown).

When all the motivational influence has been left-or percolated from the higher-level objective to the lower-level objectives, some of the lower-level objectives will be widely motivated and others will not. Only those lower-level objectives that are well-motivated will actually influence the next stage in the manufacture of the marker: the percolation process and the activation process of the interactors are not determined in a deterministic way. In this way, the target network allows a symbolic guide, to activate the objective of the manufacturing process, while arbitrarily random fluctuations in the processing are allowed simultaneously. lower level, via one or more interactors. This operation then produces some kind of measure of your satisfaction, that is, the degree to which the operation was * successful. In the second sweep from right to left through the target network, this satisfaction is passed behind the lower-level objectives to the upper-level objectives. This allows the lower level objectives to do two things: 1. notify the objective of the upper level of the overall satisfaction contained in the current state of the marker, and 2. exert a lower-superior influence on the motivation of the higher-level objective in the next stage of processing.

The satisfaction that passes behind the target network influences the motivations in the next sweep from left to right via a memory capacity in each objective.

Each objective in the target network as shown in Figure 2 potentially has two capabilities: TREND and MEMORY. TREND is the inherent tendency of an objective that is going to be either highly active (1.0) or not active (0.0), or somewhere between 5 these. Some objectives (for example, MARKER, BANDS and PANELS) are inherently not active unless they are influenced by other objectives; other objectives (eg, budgets) tend to be highly inherent assets unless they are influenced to the contrary for other purposes. TREND is the first number in brackets in each objective shown in Figure 2; the second number in brackets is the MEMORY of the objective. MEMORY is similar to inertia and indicates the non-response of an objective: it represents the degree to which an objective is resistant to influence other objectives. An objective with zero MEMORY (0.0) is totally under the influence of external pressures, while the objective with MEMORY (1.0) is not affected in any way by external influences: this "remember" your old activation status perfectly.

The above explanation allows us to understand the meaning of the objective PRESUPPOSITION in Figure 2. -The PRESUPPOSITION is highly inherent active (1.0) and has a fairly high memory (0.8). It has a disabling influence f Rete (-1.0) on the target PANELS, and this msQo in the normal course of cases ensures that the target PANELS are not active. However, any sat isf ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ This means that in the case where the BANDS are widely satisfied, the objective BUDGET is disconnected, thus allowing the PANELS to become highly motivated in the next sweep of motivation through the target network. In this way the objective PANELS can be said to presuppose the satisfaction of the BANDS.

The global target network is configured by means of a row of the target network. The following is a sample of the text of the row of the global network required to build the target network of Figure 2: 1. The MARKER uses the BANDS to extend 0.4. 2. The MARKER uses the PANELS to extend 0.6. 3. The PANELS presuppose BANDS.

Line 3, automatically adjusts an implicit antecedent objective PRESUPPOSITION that implements the relationship * r-? * of presupposition as described above. It is unheard of, important dai K &fjgfßí? Ta of the nature of this relationship: it is very different from a time sequencing relationship of the form "First make the Bands and then make the Panels". Such a relationship specifies the strict order in which things are to be done, and is highly unacceptable to the circumstance. In contrast, the presupposition relationship simply suggests a distribution of attention among other objectives at any given time: "only pay attention to the PANELS if the BANDS are highly satisfied". This relationship allows BANDS to be changed later in the marking process, handling other waiting targets until changes to BANDS have been completed. Also, the presupposition relation is not deterministic: it can be done in a strict way such that the BANDS are always completed before the PANELS, but it can also be made loose enough that the work begins in the PANELS before the BANDS are completely satisfied. .

Clusters are the central symbolic aspect of the description of the state of the current marker at any given time. They are the fundamental elements that are manipulated by the interactors, and constitute a grouping based geometrically on the installed pieces and / or the additional sub-groupings that are relatively uncoupled from the other groupings in the marker. They are characterized by their geometric around (which often overlaps with the other groupings) and by the type of pieces they allow to contain. An example of a highly restricted grouping in Figure 1 is the group of bands at the bottom right of the marker - the group is tightly restricted geographically and also in terms of the type of piece it might contain. By comparison, the rows on the leftmost side in Figure 1 are only restricted freely on their right side (the area marked by the letter a) and could contain almost all types of pieces.

As an example, the small-scale problem of the two markers in Figure 3, which consists entirely of rectangular pieces without any use of rotation or inclination, will be considered in the following.

The marker in Figure 3A was installed in a non-grouped direction, whereas in Figure 3B, it groups the bands in the lower right part of the marker, and secures a sharp defined row shown here between the pieces and K. Simply defining these groups, the average run time for this marker is reduced (for example: from 146.12 seconds to 106.3 seconds), but if the piece type restricts the system, the piece of relatively difficult pieces X or J should be placed inside from the row further to the left side, it is imposed further after the time is further reduced (for example: up to 37.15 seconds), as shown in Figure 3B.

The following concepts of interactors in a target network will be explained.

When an interactor is called in action, it performs two tasks: it operates on the marker in the same way and evaluates the results of its operation. This evaluation is then passed behind the relevant objectives in the target network. These actions of the interactor are carried out in three different stages, which are called recognition tasks or exploration, testing and construction in the following.

In the first stage, to avoid the waste of precious computing resources, an interactor simply establishes whether it should continue its activity. For this prog-deposit, the interactor is set in motion his task of recognition. The recognition task determines whether the approximate conditions for the execution of the interactor are satisfied completely.

If the acknowledgment task finds that the conditions of the interactor are not completely satisfied, then the interactor stops here. If, however, the conditions are completely satisfied, then a more extensive verification of the test task is carried out. The test task assumes that the approximate verification carried out by the recognition is correct, and the interactor can continue with its actions, but before this happens, the test task verifies what would be the benefit of continuing the activities of the interactor. The result is an assertion of the probable improvement in the score that would be expected to carry out a fully detailed calculation. The test task also usually saves several initial calculations required after construction.

Again, if the test task finds that the benefits of executing the interactor are very low, then the interactor stops here. If, however, the benefits are high enough, the interactor executes its construction task. The construction task performs all the necessary calculations to satisfy the objectives of the interactor. This could involve the allocation of computing resources. For this reason, recognition and testing tasks are performed first. When the construction task is finished, the work of the interactor is completed; Some procedure has been performed on the scoreboard and an accurate evaluation of the effects is passed behind the target network.

An important point about the interactors is that any interactor can do his work totally independent of all considerations apart from the present state of the marker. The interactors never have any particular activity that has been done previously, but instead they take their notice exclusively from patterns that look in the current state of the results in the configuration of the marker.

In the following, a typical marker making process in the case of simple rectangular pieces will be described as illustrated in Figure 3B.

^ T ^ £ * 3í ~ ** & ~ £ - * The method and system for the production of the automatic marker according to the invention, is based on the following information: - A row of definition of the marker - contains the description of the pieces and their properties. a row of constraint - contains the constraints that are relevant to the marker.

- A row of the target network - configures the objectives and interactors of the target network that codifies a particular installation strategy.

- A row of taxonomy - specifies a tax.onomy for the types of pieces, p. ex. , Panel, Watch straps is, etc.

After the appropriate selection of marker rows, their parts, constraints, and the row of the target network, which are stored in some storage medium, for example, hard disk, floppy disk, etc., they are loaded by means of processing, that is, a computer. A target network is then constructed according to the specification in the target lane array, as shown in Figure 4. The target network in Figure 4 corresponds to the following row of the target network: // Central criterion: Criterion Top (installed marker). // O j ective: Goal BandStacking (band stacking) Goal RestLaying (place the remaining pieces) Goal TierLaying (place the row on the left and left side). Goal HPLaying (remaining pieces of the HodgePodge pieces). // Keys of type: Sortkey YspanDesc: (Duration Y, descending). // Interactors: BandLayer Interactor (Band Layer): Stacker (Bands, WaistBand, *, [1, -1], 0). Tierlnitr Interactor (Tier Initiator): Tierlnitiator (Tier, l, *, YSpanDesc, left, [1,4], [0, 0, 0.05], *, *, 0.9, 0.85). Interactor HPInítr (initiator HodgePodge): Tierlnit iator (HP, *, *, YSpanDesc, left, [-1,4], [0.05, 0.1], *, Bands 81.0], 0.88, 0.85). & fr tá * J. { r5! & amp; ei! Ei * ^ * ¡ß t e e t T T ((((((((((((((((((((((((( Juggler (Tier, *,, YSpanDescend, *). // Target dependencies: Top uses BandStacking to extend 0.2 Top uses RestLaying to extend 0.8. BandStacking uses BandLayer. RestLaying presupposes BandStacking. RestLaying uses TierLaying to extend 0.4 RestLaying uses HPLaying to extend 0.6. HPLaying presupposes TierLaying. TierLaying motivates Tierlnitr. TierLaying uses TierJuggler. TierJuggler presupposes Tierlnitr. HPLaying motivates HPInitr. HPLaying uses HPJuggler. HPJuggler assumes HPInitr.

In Figure 4 the targets are indicated by circles and the metatectors are indicated by rectangles. The various parameters in the interactor specifications indicate the specific details of the interactor such as the particular installation order and the criteria for building a group.

»When the target network has been built, the top network is simply artificially motivated. This is continued until the higher level objective returns to a sufficient level of satisfaction. In this stage, the interactors have fulfilled their purpose or have been satisfied that they can no longer proceed. After the elaboration of the automatic marker is completed.

Figures 5A to 5H and 6A to 6K show some of the intermediate states of the marker during an automatic marker making process performed in accordance with the invention. In Figures 5 and 6, the rectangular pieces are indicated by the same characters. It can be seen that during the production of the automatic marker according to the invention, the intermediate markers differ slightly with each other slightly, but the parameters given subsequently improve each marker. In Figure 5, a process is shown at the end of which all the pieces are constructed, while at the end of the process shown in Figure 6 a part has not been built, although the same configuration rows have been used. "k ^" * It is noted that in relation to this feGhd I best method known by the applicant to implement the aforementioned invention, is the conventional for the manufacture of the objects to which it refers.

Having described the invention as above, the content of the following is claimed as property.

Claims (9)

CLAIMS ^^ «á-

1. The method for placing a number of pieces in a marker, characterized the stages of: a) define the restrictions (eg they are relevant in the scoreboard, b) define the pieces and their properties, define the groups in the marker by the geometric surroundings and / or by the type of pieces that are allowed to contain, d) define a hierarchy of the groups through the structure of the network, e) define one or more interactions for each lower level of groups in the hierarchy and f) perform the interactions in succession of the defined hierarchy.

2. The method according to claim 1, characterized in that the defined restriction is to minimize the wear and tear of the marker by a superimposition between the pieces.

3. The method according to claim 1, characterized in that the defined restriction is to maximize the number of pieces placed on the marker without any overlap between the pieces.

. The method according to any of claims 1 to 3, characterized in that the groups and / or interactions at the same level within the hierarchy are connected by a presupposition relation defining a time sequence relationship of these groups and / or the interactions that depend on the degree to which the corresponding interaction is achieved or passes behind the level that was successful.

5. The method according to any of claims 1 to 4, characterized in that the interactions perform the operations of placing and orienting the pieces and / or groups of pieces in a marker.

6. The method according to any of claims 1 to 5, characterized in that the pieces 2'3 of the pieces are pieces for a fabric marker consisting of wrist bands, panels, watch straps and belts.

7. The method according to claim 6, characterized in that the marker is rectangular and in the lower right part of the marker there is a group that is slightly restricted by the side and type of wrist bands.

8. The method according to claim 7, characterized in that from the left side to the right there are groups freely restricted by the size and type of the panels.

9. The storage means for storing the programming elements by means of a processing means, characterized in that the programming element comprises the coded portions for performing the steps according to one of claims 1 to 8 by means of the processing means.