METHOD FOR SEQUENCING LOADS IN AN AUTOMATED DISTRIBUTION SYSTEM, WITH REDUCTION OF DISORDER DURING A COLLECTION OF LOADS ON A COLLECTOR

Abstract

A method for sequencing loads in an automated load-distribution system having k sources with k2; at least one destination; k FIFO-type source buffer devices, each receiving loads from one of the k sources; a collector collecting the loads coming from the k source buffer devices and transporting them to the at least one destination. The collector has k successive nodes each collecting the loads coming from one of the source buffer devices. The control system processes customer orders listing loads for a given destination and being associated with a sequential order number of destination. The control system: builds a collection list containing n loads to be collected and reducing a disorder of the n loads relative to a rising order of the sequential order numbers of destination; and controlling the collector and the source buffer devices to collect loads on the collector in compliance with the collection list.

Claims

1. A method for sequencing loads, implemented by a control system in an automated load-distribution system comprising: k sources with k2; at least one destination; k source buffer devices of a first-in first-out type, each receiving loads coming from one of the k sources; a collector collecting the loads coming out of the k source buffer devices and transporting them to said at least one destination, the collector comprising k successive nodes each configured to collect the loads coming out of one of the source buffer devices; and said control system configured to process customer orders, each customer order listing loads for a given destination and being each associated with a sequential order number of destination; wherein the method comprises the following acts performed by the control system: building a collection list containing n loads to be collected and reducing a disorder of said n loads relative to a rising order of the sequential order numbers of destination, said n loads being contained in the source buffer devices, n=.sub.i=1.sup.i=kp(i), with p(i) being a number of loads to be collected in the i.sup.th source buffer device; and controlling the collector and the source buffer devices to carry out a collection of loads on the collector in compliance with said collection list, wherein the act of building the collection list comprises the following acts: (a) initializing a first set of states E1 with a single state e.sub.init=(U.sub.init, L.sub.init), where U.sub.init is a k-uplet containing k zeros and L.sub.init is an empty list; (b) initializing a second set of states E2 with an empty value; (c) for each building act among n successive building acts: (c.1) for each state e of E1, with e=(U, L), where U is a k-uplet containing k elements, U=(z.sub.1, . . . , z.sub.k) with z.sub.i a number of loads taken from the i.sup.th source buffer device, i{1, . . . , k}, and L is a list of loads associated with U: (c.1.1) for each value of f{1, . . . , k}: (c.1.1.1) if U(f)<p(f), with U(f) being a number of loads of the f.sup.th source buffer device contained in L, and p(f) the number of loads to be collected in the f.sup.th source buffer device: i creating a new state e.sub.N=(U.sub.N,L.sub.N) starting from e=(U, L), in adding 1 to U(f) and in adding, at the end of L, the load occupying the (U(f)+1).sup.th position in the sequence of loads contained in the f.sup.th source buffer device; ii computing a value of disorder d of the list L.sub.N of the new state e.sub.N, with a function of computation of disorder relative to a rising order of the sequential order numbers of destination; iii if E2 contains another new state e.sub.N=(U.sub.N,L.sub.N), with U.sub.N=U.sub.N and d a value of disorder of the list L.sub.N: if d<d, e.sub.N replaces e.sub.N in E2, and if dd, e.sub.N is not inserted into E2; iv if E2 does not contain said other new state e.sub.N, e.sub.N is inserted into E2; (c.2) if the building act is not the n.sup.th building act: E2 becomes the new set of states E1 and the method passes to the next building act; (c.3) if the building act is the n.sup.th building act: E2 contains a single final state e.sub.F=(U.sub.F,L.sub.F) and L.sub.F forms said collection list.

2. The method according to claim 1, wherein, at the act (c.1.1.1), the control system also verifies whether N(f)<y.sub.f, with N(f) being the longest sequence of loads of the f.sup.th source buffer device placed consecutively in L, and y.sub.f being a predetermined threshold, and wherein the acts (i) to (iv) are carried out only if U(f)<p(f) and N(f)<y.sub.f.

3. The method according to claim 1, wherein the act ii is followed by the following act: iia if d>d.sub.H, with d.sub.H being a predetermined value, the new state e.sub.1 is not inserted into E2 and the acts (iii) and (iv) are not carried out.

4. The method according to claim 3, comprising computing the predetermined value d.sub.H as follows: building a reference list L.sub.H containing said n loads and built as follows: the first load of L.sub.H is a load having the smallest sequential order number of destination among the loads truly ready to go out of the k source buffer devices; and each following load of L.sub.H is the load having the smallest sequential order number of destination among the loads that are ready to go out of the k source buffer devices, assuming fictitiously that the preceding loads of L.sub.H have come out of the k source buffer devices; and computing d.sub.H, as a value of disorder of the list L.sub.H, with said disorder computing function.

5. The method according to claim 1, wherein said disorder computing function, for a list M of q loads, is written: H(M)=.sub.i=1.sup.i=q[(i1)K(i)], with K(i) being the number of loads of the list M placed before the i.sup.th load of the list M and having a sequential order number of destination smaller than or equal to the sequential order number of the i.sup.th load of the list M.

6. The method according to claim 1, wherein the loads of a given customer order must reach a given destination in a given rising sequential order of destination, and wherein said control system carries out an act of controlling at least one sequencing device, placed between the collector and said at least one destination, to make a correction of the disorder of the n loads.

7. The method according to claim 1, wherein the control system performs the following act, before the act of building the collection list, for at least one group of R successive loads contained in one of the source buffer devices, with R being an integer greater than or equal to 2: computing a substitute sequential order number of destination as a function of the sequential order numbers of destination of the R loads; and wherein, for the execution of the act of building the collection list, the control system uses the substitute sequential order number of destination for each of the R loads.

8. The method according to claim 7, wherein the computation of a substitute sequential order number of destination as a function of the sequential order numbers of destination of the R loads comprises: computing an average value of the sequential order numbers of destination of the R loads; computing a value of disorder of the R loads as a function of the sequential order numbers of destination of the R loads; if the value of disorder of the R loads is above a predetermined threshold, the substitute sequential order number of destination is the average value rounded up to the next integer; if not, the substitute sequential order number of destination is an average value rounded down to the next integer.

9. The method according to claim 1, wherein a new execution of the acts of the method is launched if an entry of at least one new load into one of the source buffer devices prompts a modification of the loads to be collected in said source buffer device and therefore of the n loads to be collected in the set of the k source buffer devices.

10. (canceled)

11. A non-transitory computer-readable storage medium storing a computer program comprising program code instructions for implementing a method for sequencing loads by a control system in an automated load-distribution system, when said program is executed on a computer of the control system, wherein: the control system comprises: k sources with k2; at least one destination; k source buffer devices of a first-in first-out type, each receiving loads coming from one of the k sources; a collector collecting the loads coming out of the k source buffer devices and transporting them to said at least one destination, the collector comprising k successive nodes each configured to collect the loads coming out of one of the source buffer devices; and said control system configured to process customer orders, each customer order listing loads for a given destination and being each associated with a sequential order number of destination; and wherein the instructions configure the control system to perform acts comprising: building a collection list containing n loads to be collected and reducing a disorder of said n loads relative to a rising order of the sequential order numbers of destination, said n loads being contained in the source buffer devices, n=.sub.i=1.sup.i=kp(i), with p(i) being a number of loads to be collected in the i.sup.th source buffer device; and controlling the collector and the source buffer devices to carry out a collection of loads on the collector in compliance with said collection list, wherein the act of building the collection list comprises the following acts: (a) initializing a first set of states E1 with a single state e.sub.init=(U.sub.init, L.sub.init), where U.sub.init is a k-uplet containing k zeros and L.sub.init is an empty list; (b) initializing a second set of states E2 with an empty value; (c) for each building act among n successive building acts: (c.1) for each state e of E1, with e=(U, L), where U is a k-uplet containing k elements, U=(z.sub.1, . . . , z.sub.k) with z.sub.i a number of loads taken from the i.sup.th source buffer device, i{1, . . . , k}, and L is a list of loads associated with U: (c.1.1) for each value of f{1, . . . , k}: (c.1.1.1) if U(f)<p(f), with U(f) being a number of loads of the f.sup.th source buffer device contained in L, and p(f) the number of loads to be collected in the f.sup.th source buffer device: i creating a new state e.sub.N=(U.sub.N,L.sub.N) starting from e=(U, L), in adding 1 to U(f) and in adding, at the end of L, the load occupying the (U(f)+1).sup.th position in the sequence of loads contained in the f.sup.th source buffer device; ii computing a value of disorder d of the list L.sub.N of the new state e.sub.N, with a function of computation of disorder relative to a rising order of the sequential order numbers of destination; iii if E2 contains another new state e.sub.N=(U.sub.N,L.sub.N), with U.sub.N=U.sub.N and d a value of disorder of the list L.sub.N: if d<d, e.sub.N replaces e.sub.N in E2, and if d>d, e.sub.N is not inserted into E2; iv if E2 does not contain said other new state e.sub.N, e.sub.N is inserted into E2; (c.2) if the building act is not the n.sup.th building act: E2 becomes the new set of states E1 and the method passes to the next building act; (c.3) if the building act is the n.sup.th building act: E2 contains a single final state e.sub.F=(U.sub.F,L.sub.F) and L.sub.f forms said collection list.

12. Automated load distribution system comprising: k sources with k2; at least one destination; k source buffer devices, of a first-in first-out type, each receiving loads coming from one of the k sources; a collector collecting loads going out of the k source buffer devices and transporting them towards said at least one destination, the collector comprising k successive nodes each configured to collect the loads going out of one of the k source buffer devices; and a control system configured to process customer orders, each customer order listing loads for a given destination and each being associated with a sequential order number of destination; wherein said control system comprises a computation machine configured to: build a collection list containing n loads to be collected and reducing a disorder of said n loads relative to a rising order of the sequential order numbers of destination, said n loads being contained in the source buffer devices, with n=.sub.i=1.sup.i=kp(i), p(i) being a number of loads to be collected in the i.sup.th source buffer device; and control the collector and the source buffer devices for a collection of loads on the collector compliant with said collection list; and wherein the computation machine is configured to build the collection list in carrying out the following acts: (a) initializing a first set of states E1 with a single state e.sub.init=(U.sub.init, L.sub.init), where U.sub.init is a k-uplet containing k zeros and L.sub.init is an empty list; (b) initializing a second set of states E2 with an empty value; (c) for each building act among n successive building acts: (c.1) for each state e of E1, with e=(U, L), where U is a k-uplet containing k elements, U=(z.sub.1, . . . , z.sub.k) with z.sub.i a number of loads taken from the i.sup.th source buffer device, i{1, . . . , k}, and L is a list of loads associated with U: (c.1.1) for each value of f{1, . . . , k}: (c.1.1.1) if U(f)<p(f), with U(f) being a number of loads of the f.sup.th source buffer device contained in L, and p(f) is the number of loads to be collected in the f.sup.th source buffer device: i creation of a new state e.sub.N=(U.sub.N,L.sub.N) starting from e=(U, L), in adding 1 to U(f) and in adding, at the end of L, the load occupying the (U(f)+1).sup.th position in the sequence of loads contained in the f.sup.th source buffer device; ii computing a value of disorder d of the list L.sub.N of the new state e.sub.N, with a function of computation of disorder relative to a rising order of the sequential order numbers of destination; iii if E2 contains another new state e.sub.N=(U.sub.N,L.sub.N), with U.sub.N=U.sub.N and d a value of disorder of the list L.sub.N: if d<d, e.sub.N replaces e.sub.N in E2, and if dd, e.sub.N is not inserted into E2; iv if E2 does not contain said other new state e.sub.N, e.sub.N is inserted into E2; (c.2) if the building act is not the n.sup.th building act: E2 becomes the new set of states E1 and the method passes to the next building act; (c.3) if the building act is the n.sup.th building act: E2 contains a single final state e.sub.F=(U.sub.F,L.sub.F) and L.sub.F forms said collection list.

Description

4. LIST OF FIGURES

[0104] Other features and advantages of the invention shall appear from the following description, given by way of an indicative and non-exhaustive example and from the appended drawings of which:

[0105] FIGS. 1A, 1B and 1C, already described with reference to the prior art, illustrate three states (loads before collection on the collector, loads after collection on the collector, and loads arriving at the picking or preparing station after coming out of the collector) of the processing of a customer order with the first known solution;

[0106] FIG. 2 presents a block diagram of an example of an automated distribution system in which it is possible to implement a load-sequencing method according to the invention;

[0107] FIG. 3 presents a flowchart of a load-sequencing method according to one particular embodiment of the invention;

[0108] FIG. 4 presents the structure of a control system according to one particular embodiment of the invention;

[0109] FIG. 5 illustrates an algorithm of a particular implementation of the step 31 of FIG. 3 (building of the collection or picking list);

[0110] FIG. 6 illustrates an example of a state tree browsed by the algorithm of FIG. 5;

[0111] FIGS. 7A, 7B and 7C illustrate three states (loads before collection on the collector, loads after collection on the collector and loads after final scheduling) of the processing of a customer order with a load-sequencing method according to the invention, in a first implementation with a single destination;

[0112] FIGS. 8A, 8B and 8C illustrate three states (loads before collection on the collector, loads after collection on the collector and loads after final scheduling) of the processing of a customer order with a load-sequencing method according to the invention, in a second implementation with several destinations;

[0113] FIGS. 9A and 9B illustrate two states (loads before collection on the collector and loads after collection on the collector) of the processing of a customer order with a load-sequencing method according to the invention, in a third implementation with a use of substitute sequential order numbers of destination.

5. DETAILED DESCRIPTION

[0114] In all the figures of the present invention, the identical elements and steps are designated by same numerical references.

[0115] FIG. 2 presents a block diagram of an example of an automated distribution system in which it is possible to implement a load-sequencing method according to the invention. In this example, the system comprises five source S1 to S5 (for example, different parts (storage units) of a storage depot), five destinations D1 to D5 (for example customer order picking or preparing stations), a collector 1 (composed for example of one or more conveyors), sequencing and storage buffer devices or systems 91 and a control system 90 (for example of the WCS type). In this example, the number of sources and the number of destinations are given purely as an illustration. More generally, the system comprises k sources with k2, and at least one destination.

[0116] As already explained further above, the collector 1 is configured to transport loads up to each destination and comprises a plurality of successive nodes: those referenced N1 to N5 are each configured to collect loads coming out of one of the sources S1 to S5 and those referenced N1 to N5 are each configured to direct loads towards one of the destinations D1 to D5. Each of these nodes comprises for example a 90-degree or 45-degree transfer device.

[0117] Each of the sources S1 to S5 is for example connected to one of the nodes N1 to N5 by a FIFO type source buffer device F1 to F5. Similarly, each of the destinations D1 to D5 is for example connected to one of the nodes N1 to N5 by a FIFO type destination buffer device F1 to F5.

[0118] Upstream to each destination, a sequencing and buffer storage system 91 enables a final scheduling of the loads in a rising sequential order of destination for this destination. As described in detail further below, it is accepted that, at the end of the collection of loads on the collector, these loads are in disorder (relative to the rising sequential order of destination). The sequencing and buffer storage system 91 eliminates this disorder.

[0119] In one variant, the constraint is more flexible as regards the destinations and it is accepted that the rising order of sequential order numbers of destination is not complied with by the loads arriving at this destination. In this variant, the sequencing and buffer storage system 91 upstream to each destination is either omitted (not present) or used to carry out a final scheduling that can only be partial (i.e. that sometimes only reduces the above-mentioned disorder without eliminating it).

[0120] In another variant, there are not several sequencing and buffer storage systems 91 (one just upstream to each destination and downstream to the collector 1) but only one sequencing and buffer storage system 91 (upstream to the set of destinations).

[0121] The control system 90 is configured to process customer orders each listing loads to be extracted from the sources and ideally (see discussion here above) to be provided, in a given rising sequential order of destination, to a given destination.

[0122] For example, it implements the load-sequencing method according to the particular embodiment described here below with reference to FIGS. 5 and 6.

[0123] Referring now to FIG. 3, we present a load-sequencing method according to one particular embodiment of the invention. This method is implemented by the control system 90.

[0124] In a step 31, the control system 90 prepares a collection list containing n loads to be collected and reducing a disorder of the n loads relative to a rising order of sequential order numbers of destination. The n loads are contained in the source buffer devices F1 to F5. We have: n=.sub.i=1.sup.i=kp(i), with p(i) being a number of loads to be collected in the i.sup.th source buffer device. One particular implementation of this step 31 for building the collection list is described here below with reference to FIG. 5.

[0125] In a step 32, the control system 90 controls the collector 1 and the source buffer devices F1 to F5 so that a collection of loads (on the collector) is carried out in compliance with the collection list.

[0126] If the loads of a given customer order must arrive at a given destination in a given rising sequential order of destination, a step 33 is carried out in which the control system 90 controls the sequencing and buffer storage systems 91 for a correction of disorder of the n loads.

[0127] In a test step 34, the control system 90 verifies that an entry of at least one new load into one of the source buffer devices F1 to F5 prompts a modification of the loads to be collected in this source buffer device and therefore a modification of the n loads to be collected in the set of k source buffer devices. In the event of a positive response at the test step 34 (i.e. in the event of a modification of the set of n loads to be collected), the control system 90 launches a new execution of the steps of the method.

[0128] In one variant, the load-sequencing method comprises a preliminary step 30 which is described further below with reference to FIGS. 9A and 9B.

[0129] FIG. 4 presents the structure of a control system 90 according to one particular embodiment of the invention. This control system comprises a random-access memory 92 (for example a RAM), a processing unit 91, equipped for example with a processor and driven by a computer program 930 stored in a read-only memory 93 (for example a ROM or a hard disk drive).

[0130] At initialization, the code instructions of the computer program are for example loaded into the random-access memory 92 and then executed by the processor of the processing unit 91 to implement the load-sequencing method of the invention (for example according to the embodiment of FIGS. 3 and 5). The processing unit 91 inputs commands (also referred to as instructions) 94. The processor of the processing unit 91 processes these commands 94 and generates other commands (also referred to as instructions) 95 at output enabling the control (or commanding) of the different elements comprised in the automated distribution system, especially the sources S1 to S5, the FIFO-type source buffer devices F1 to F5, the collector 1, the destination D1 to D5, the FIFO-type destination buffer devices F1 to F5 and the sequencing and buffer storage systems 91.

[0131] This FIG. 4 illustrates only one particular way, among several possible ways, of carrying out the technique of the invention in any one of its embodiments. Indeed, the control system can be carried out equally well on a reprogrammable computing machine (for example a PC computer, a DSP processor, a microcontroller etc.) executing a program comprising a sequence of instructions, or on a dedicated computation machine (for example a set of logic gates such as an FPGA or an ASIC or any other hardware module).

[0132] Should the control system be made with a reprogrammable computing machine, the corresponding program (i.e. the sequence of instructions) could be stored in a storage medium that is detachable (such as for example a floppy disk, a CD-ROM or a DVD-ROM) or non-detachable, this storage medium being partially or totally readable by a computer or a processor.

[0133] FIG. 5 illustrates an algorithm of a particular implementation of the step 31 of FIG. 3 (preparing the collection list LC).

[0134] In a step 501, the control system initializes a first set of states E1 with a single state e.sub.init=(U.sub.init, L.sub.init), where U.sub.init is a k-uplet containing k zeros and is an empty list.

[0135] In a step 502, the control system initializes a second set of states E2 with a zero value.

[0136] In a test step 503, the control system verifies whether n successive building steps (i.e. all of them) have been carried out.

[0137] In the event of a positive response at the test step 503, the control system passes to the step 516 in which it obtains the collection list LC from a single final state e.sub.F=(U.sub.F,L.sub.F) contained in E2. Indeed, it takes L.sub.F as a collection list LC.

[0138] In the event of a negative response at the testing step 503, the control system starts the processing of the next building step in passing to the test step 504 in which it verifies whether all the states of E1 have been processed. Each state e of E1 is written as e=(U, L), where U is a k-uplet containing k elements, U=(z.sub.i, z.sub.k) with z, as a number of loads taken in the i.sup.th source buffer device, i{1, . . . , k}, and L is a list of loads associated with U.

[0139] In the event of a positive response at the test step 504, the control system passes to the step 515 in which, if the building step is not the n.sup.th building step, E2 becomes the new set of states E1, and then the control system returns to the step 503 (for the passage to the next building step).

[0140] In the event of a negative response at the test step 504, the control system takes an unprocessed state E1 and passes to the test step 505 in which it verifies whether all the values off have been processed with f{1, . . . , k}.

[0141] In the event of a positive response at the step 505, the control system returns to the step 504. In the event of a negative response to the test step 505, the control system takes an unprocessed value of f and passes to the test step 506 in which it verifies whether U(f)<p(f), with U(f) being a number of loads of the f.sup.th source buffer device contained in L, and p(f) being the number of loads to be collected in the f.sup.th source buffer device.

[0142] In the event of a negative response at the test step 506, the control system returns to the step 505. In the event of a positive response at the step 506, the control system passes to the test step 506a in which it verifies whether N(f)<y.sub.f, with N(f) being a maximum number of loads of the f.sup.th source buffer device placed consecutively in L, and y.sub.f a predetermined value (for example, y.sub.f=6).

[0143] In the event of a negative response at the test step 506a, the control system returns to the step 505. In the event of a positive response at the test step 506a, the control system passes to the step 507 in which it creates a new state e.sub.N=(U.sub.N,L.sub.N) starting from e=(U, L), in adding 1 to U(f) and adding, at the end of L, the load occupying the (U(f)+1).sup.th position in the sequence of loads contained in the f.sup.th source buffer device.

[0144] The step 507 is followed by the step 508 in which the control system computes a value of disorder d of the list L.sub.N of the new state e.sub.N, with a function of computation of disorder relative to a rising order of sequential order numbers of destination.

[0145] In one particular embodiment of the step 508, the control system uses a function of computation of disorder which, for a list M of q loads, is written as follows:

H(M)=.sub.i=1.sup.i=q[(i1)K(i)][Equation 1]

[0146] with K(i) being the number of loads of the list M placed before the i.sup.th load of the list M and having a sequential order number of destination smaller than or equal to the sequential order number of the i.sup.th load of the list M.

[0147] Other disorder computation functions can be used without departing from the framework of the present invention, especially but not exclusively:

B(M)=MAX(K(i)),i{1, . . . ,q}[Equation 2]

[0148] with K(i) as defined further above.

F(M)=.sub.i=1.sup.i=q|(A(i)i|[Equation 3]

[0149] with A(i) being the position that would be occupied by the i.sup.th load of the list M if the q loads of the list M were re-ordered according to a rising order of sequential order numbers of destination.

G(M)=MAX(|(A(i)i|),i{1, . . . ,q}[Equation 4]

[0150] with A(i) as defined further above.

[0151] For example, with M=(3,1,8,4,7,2,6,5), we obtain: [0152] H(M)=([(11)0]+[(21)0]+[(31)2]+[(41)2]+[(51)3]+[(61)1]+[(71)4]+[(81)4])=12 [0153] B(M)=4 [0154] F(M)=(|31|+|12|+|83|+|44|+|75|+|26|+|67|+|58|)=18 [0155] G(M)=5

[0156] The step 508 is followed by the test step 509 in which the control system verifies whether d>d.sub.H, with d.sub.H being a predetermined value.

[0157] In one particular embodiment of the step 509, the predetermined value d.sub.H is computed as follows: [0158] building a reference list L.sub.H that contains the n loads and is built as follows: [0159] the first load of L.sub.H is the load having the smallest sequential order number of destination among the loads really ready to go out of the k source buffer devices; [0160] each following load of L.sub.H is the load having the smallest sequential order number of destination among the loads that would be ready to go out of the k source buffer devices in fictitiously assuming that the previous loads of L.sub.H have gone out of the k source buffer devices; [0161] computing d.sub.H as a value of disorder of the list L.sub.H, with the disorder computation function.

[0162] In the event of a positive response at the test step 509, the control system returns to the step 505. In the event of a negative response at the test step 509, the control system goes to the test step 510 in which it verifies whether E2 contains another new state e.sub.N=(U.sub.N,L.sub.N), with U.sub.N=U.sub.N and d being a value of disorder of the list L.sub.N.

[0163] In the event of a positive response at the test step 510, the control system goes to the step 512 in which it verifies whether d<d. In the event of a positive response at the test step 512, the control system goes to the step 514 in which it replaces e.sub.N by e.sub.N in E2. In the event of a negative response at the test step 512, the control system goes to the step 513 in which it does not insert e.sub.N into E2. At the end of the step 512 or the step 514, the control system returns to the step 505.

[0164] In the event of a negative response at the test step 510, the control system goes to the step 511 in which it inserts e.sub.N into E2, and then returns to the step 505.

[0165] Referring now to FIG. 6 and FIGS. 7A, 7B and 7C, we present the processing of a customer order with a load-sequencing method according to the invention, in a first implementation with only one destination. FIGS. 7A, 7B and 7C respectively illustrate the following three steps of this processing: loads before collection on the collector, loads after collection on the collector and loads after final scheduling.

[0166] In this example, it is assumed that there are two sources S1 and S2 and one destination D1. The customer order to be processed lists the following loads: 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10 (in the figures the loads are referenced by their sequential order number of destination).

[0167] As illustrated in FIG. 7A, the source buffer device F1 (downstream from the source S1) contains five loads according to the following sequence: 7, 6, 1, 9 and 3. The source buffer device F2 (downstream from the source S2) contains five loads according to the following sequence: 8, 4, 5, 2 and 10. We therefore have, n=10, p(1)=5 and p(2)=5.

[0168] The results of the step for initializing E1 and of the n building steps described further above with reference to FIG. 5 are the following (in taking H(M) as a function of computation of the value of disorder d):

[0169] Step for Initializing E1

[0170] state e.sub.00=[U=(0,0); L=( )]; d=0

[0171] Building Step 1

[0172] state e.sub.10=[U=(1,0) L=(8)]; d=0

[0173] state e.sub.01=[U=(0,1) L=(7)]; d=0

[0174] Building Step 2

[0175] state e.sub.20=[U=(2,0) L=(8,4)]; d=1

[0176] state e.sub.n=[U=(1,1) L=(7,8)]; d=0

[0177] state e.sub.11=[U=(1,1); L=(8,7)]; d=1 (not kept)

[0178] state e.sub.02=[U=(0,2) L=(7,6)]; d=1

[0179] Building Step 3

[0180] state e.sub.30=[U=(3,0) L=(8,4,5)]; d=2

[0181] state e.sub.21=[U=(2,1) L=(8,4,7)]; d=2

[0182] state e.sub.21=[U=(2,1); L=(7,8,4)]; d=2 (not kept)

[0183] state e.sub.12=[U=(1,2); L=(7,8,6)]; d=2 (not kept)

[0184] state e.sub.12=[U=(1,2); L=(7,6,8)]; d=1

[0185] state e.sub.03=[U=(0,3); L=(7,6,1)]; d=3

[0186] Building Step 4

[0187] state e.sub.40=[U=(4,0); L=(8,4,5,2)]; d=5

[0188] state e.sub.31=[U=(3,1); L=(8,4,5,7)]; d=3

[0189] state e.sub.31=[U=(3,1); L=(8,4,7,5)]; d=4 (not kept)

[0190] state e.sub.22=[U=(2,2); L=(8,4,7,6)]; d=4

[0191] state e.sub.22=[U=(2,2); L=(7,6,8,4)]; d=4 (not kept)

[0192] state e.sub.13=[U=(1,3); L=(7,6,8,1)]; d=4 (not kept)

[0193] state e.sub.13=[U=(1,3); L=(7,6,1,8)]; d=3

[0194] state e.sub.04=[U=(0,4); L=(7,6,1,9)]; d=3

[0195] Building Step 5

[0196] state e.sub.50=[U=(5,0); L=(8,4,5,2,10)]; d=5

[0197] state e.sub.41=[U=(4,1); L=(8,4,5,2,7)]; d=6

[0198] state e.sub.41=[U=(4,1); L=(8,4,5,7,2)]; d=7 (not kept)

[0199] state e.sub.32=[U=(3,2); L=(8,4,5,7,6)]; d=5

[0200] state e.sub.32=[U=(3,2); L=(8,4,7,6,5)]; d=7 (not kept)

[0201] state e.sub.23=[U=(2,3); L=(8,4,7,6,1)]; d=8 (not kept)

[0202] state e.sub.23=[U=(2,3); L=(7,6,1,8,4)]; d=6

[0203] state e.sub.14=[U=(1,4); L=(7,6,1,8,9)]; d=3

[0204] state e.sub.14=[U=(1,4); L=(7,6,1,9,8)]; d=4 (not kept)

[0205] state e.sub.05=[U=(0,5); L=(7,6,1,9,3)]; d=6

[0206] Building Step 6

[0207] state e.sub.51=[U=(5,1); L=(8,4,5,2,10,7)]; d=7 (not kept)

[0208] state e.sub.51=[U=(5,1); L=(8,4,5,2,7,10)]; d=6

[0209] state e.sub.42=[U=(4,2); L=(8,4,5,2,7,6)]; d=8

[0210] state e.sub.42=[U=(4,2); L=(8,4,5,7,6,2)]; d=10 (not kept)

[0211] state e.sub.33=[U=(3,3); L=(8,4,5,7,6,1)]; d=10 (not kept)

[0212] state e.sub.33=[U=(3,3); L=(7,6,1,8,4,5)]; d=9

[0213] state e.sub.24=[U=(2,4); L=(7,6,1,8,4,9)]; d=6

[0214] state e.sub.24=[U=(2,4); L=(7,6,1,8,9,4)]; d=7 (not kept)

[0215] state e.sub.15=[U=(1,5); L=(7,6,1,8,9,3]; d=7

[0216] state e.sub.15=[U=(0,5); L=(7,6,1,9,3,8)]; d=7 (not kept)

[0217] Building Step 7

[0218] state e.sub.52=[U=(5,2); L=(8,4,5,2,7,10,6)]; d=9 (not kept)

[0219] state e.sub.52=[U=(5,2); L=(8,4,5,2,7,6,10)]; d=8

[0220] state e.sub.43=[U=(4,3); L=(8,4,5,2,7,6,1)]; d=14

[0221] state e.sub.43=[U=(4,3); L=(7,6,1,8,4,5,2)]; d=14 (not kept)

[0222] state e.sub.34=[U=(3,4); L=(7,6,1,8,4,5,9)]; d=9

[0223] state e.sub.34=[U=(3,4); L=(7,6,1,8,4,9,5)]; d=10 (not kept)

[0224] state e.sub.25=[U=(2,5); L=(7,6,1,8,4,9,3)]; d=11

[0225] state e.sub.25=[U=(2,5); L=(7,6,1,8,9,3,4)]; d=11 (not kept)

[0226] Building Step 8

[0227] state e.sub.53=[U=(5,3); L=(8,4,5,2,7,6,10,1)]; d=15 (not kept)

[0228] state e.sub.53=[U=(5,3); L=(8,4,5,2,7,6,1,10)]; d=14

[0229] state e.sub.44=[U=(4,4); L=(8,4,5,2,7,6,1,9)]; d=14

[0230] state e.sub.44=[U=(4,4); L=(7,6,1,8,4,5,9,2)]; d=15 (not kept)

[0231] state e.sub.35=[U=(3,5); L=(7,6,1,8,4,5,9,3)]; d=15

[0232] state e.sub.35=[U=(3,5); L=(7,6,1,8,4,9,3,5)]; d=15 (not kept)

[0233] Building Step 9

[0234] state e.sub.54=[U=(5,4); L=(8,4,5,2,7,6,1,10,9)]; d=15 (not kept)

[0235] state e.sub.54=[U=(5,4); L=(8,4,5,2,7,6,1,9,10)]; d=14

[0236] state e.sub.45=[U=(4,5); L=(8,4,5,2,7,6,1,9,3)]; d=20

[0237] state e.sub.45=[U=(4,5); L=(7,6,1,8,4,5,9,3,2)]; d=22 (not kept)

[0238] Building Step 10

[0239] state e.sub.55=[U=(5,5); L=(8,4,5,2,7,6,1,9,10,3)]; d=21 (not kept)

[0240] state e.sub.55=[U=(5,5); L=(8,4,5,2,7,6,1,9,3,10)]; d=20

[0241] The state e.sub.55 is therefore the final state, of which the list L=(8,4,5,2,7,6,1,9,3,10) is taken as a collection list LC. This is illustrated in FIG. 7B. After the list is re-ordered order (final scheduling) by the sequencing and buffer storage system 91, we obtain the list (1,2,3,4,5,6,7,8,9,10), as illustrated in FIG. 7C.

[0242] FIG. 6 illustrates the tree of states (and more precisely of the values of U of these states) browsed in this example by the algorithm of FIG. 5.

[0243] Referring now to FIGS. 8A, 8B and 8C, we present the processing of a customer order with a load-sequencing method according to the invention, in a second implementation with several destinations. FIGS. 8A, 8B and 8C respectively illustrate the following three states of this processing: loads before collection on the collector, loads after collection on the collector and loads after final scheduling.

[0244] In this example, it is assumed that there are three sources S1, S2 and S3 and four destinations D1, D2, D3 and D4. There are two customer orders to be processed, one for each of the destinations. Each of these customer orders lists four loads having the following sequential order numbers of destination: 1, 2, 3 and 4. In the figures, the loads are referenced by their sequential order number of destination as well as by a geometrical code corresponding to their destination (oval for D1, triangle for D2, rectangle for D3 and circle for D4).

[0245] As illustrated in FIG. 8A, the source buffer device F1 (downstream from the source S1) contains five loads according to the following sequence:3r, 1o, 2c, 4t and 1r (where the letters o, t, r and c respectively correspond to oval, triangle, rectangle and circle, to indicate their destination (D1, D2, D3 and D4 respectively). The source buffer device F2 (downstream from the source S2) contains six loads according to the following sequence: 2o, 1t, 4r, 3c, 3t and 4c. The source buffer device F3 (downstream from the source S3) contains five loads in the following sequence: 1c, 2r, 4o, 2t and 3o. We therefore have, n=16, p(1)=5, p(2)=6 and p(3)=5.

[0246] FIG. 8B illustrates the collection list obtained by applying the algorithm of FIG. 5: LC=(1c, 2o, 1t, 2r, 3r, 1o, 2c, 4t, 1r, 4o, 2t, 3o, 4r, 3c, 3t, 4c). A disorder is noted because of the positioning of the loads 1r and 3o.

[0247] FIG. 8C illustrates the loads sequenced accurately in the different destination buffer devices F1 to F5, after re-ordering (final scheduling) by the sequencing and buffer storage systems 91.

[0248] Referring now to FIGS. 9A and 9B, we present the processing of a customer order with a load-sequencing method according to the invention in a third implementation with use of substitute sequential order numbers of destination. FIGS. 9A and 9B respectively illustrate the two following states of this processing: loads before collection on the collector and loads after collection on the collector.

[0249] We take the case (the variant referred further above) where the load-sequencing method of FIG. 3 comprises a preliminary step 30 (before the step 31 for building the collection list). In this preliminary step 30, the control system computes the substitute sequential order numbers of destination as a function of the sequential order numbers of destination of the loads contained in the source buffer devices F1 to F5. More specifically, for at least one group (or each group in one particular embodiment) of R successive loads contained in one of the source buffer devices with R as an integer greater than or equal to 2: the system computes a substitute order number of destination as a function of the sequential order numbers of destination of the R loads. Then, during the execution of the step 31 for building the collection list, the control system uses the substitute sequential order number of destination for each of the R loads.

[0250] In one particular embodiment, the computation made at the step 30 comprises the following for each group of R loads: [0251] computing an average value of the sequential order numbers of destination of the R loads; [0252] computing a value of disorder of the R loads as a function of the sequential order numbers of destination of the R loads (examples of disorder computation functions have already been described further above); [0253] if the value of disorder of the R loads is greater than a predetermined threshold S, the substitute sequential order number of destination is the average value rounded up to the next integer; if not the substitute sequential order number of destination is the average value rounded down to the next integer.

[0254] The predetermined threshold is for example: S=0. In this case, we take the average value rounded down to the next integer only if there is no disorder in the R loads. In one variant, S is greater than zero (for example S=4). In this case we accept a tolerance value that takes the average value rounded down to the lower integer so long as the disorder in the R loads is smaller than S.

[0255] In the example illustrated in FIG. 9A, it is assumed that, in the example of FIG. 8A, there are three sources 51, S2 and S3 (and therefore three source buffer devices F1, F2 and F3), four destinations D1, D2, D3 and D4 and four customer orders to be processed (one for each of the destinations). We add the following constraint: the loads of the source buffer devices must be collected R by R (i.e. R at a time) on the collector 1 (with R=2 in this example).

[0256] In the example of FIG. 9A, we have the following groups (demarcated by dashes): for the source buffer device F1, (3r, 1o) and (2c,4t) (since the load 1r is alone, it is not taken into account); for the source buffer device F2, (2o, 1t), (4r,3c) and (3t,4c); for the source buffer device F3, (1c,2r) and (4o,2t) (since the load 3o is alone, it is not taken into account). In this example, the loads 1r and 3o are not part of a group of loads. Therefore for them there is no computation of a substitute order number of destination. They are not taken into account in the step 31 for building the collection list (in other words, they are not part of the n loads to be collected). They will be taken into account when the control system 90 launches a new execution of the steps of the method in the event of modification of the set of n loads to be collected (i.e. in the event of a positive response at the test step 34). In one variant, or else if there is no modification of the set of n loads to be collected, the loads 1r and 3o (with their sequential order number of destination) are immediately taken into account in the step 31 for building the collection list.

[0257] For each of these groups, the result of the computation of the step 30 (taking S=0) is indicated between brackets to the right of each load of the group. Let us consider two examples with different rounded out values: [0258] for the group (2o, 1t) of the source buffer device F2: the substitute sequential order number of destination is 2 (the average of 2 and 1=1.5; value rounded up to the next integer because there is disorder in the two loads); [0259] for the group (1c,2r) of the source buffer device F3: the substitute sequential order number of destination is 1 (the average of 1 and 2=1.5; value rounded down to the next integer because there is no disorder in the two loads).

[0260] FIG. 9B illustrates the collection list obtained by applying the algorithm of FIG. 5 with the substitute sequential order numbers of destination (except for the loads 1r and 3o which are taken into account with their sequential order number of destination): LC=(1c, 2r, 2o, 1t, 3r, 1o, 2c, 4t, 1r, 4o, 2t, 3o, 4r, 3c, 3t, 4c).