EP0508094A1 - Method for preventing local bunching of cars in an elevator group under variable traffic flow - Google Patents

Abstract

This method minimises the local accumulation of lift cabins (cluster formation) in that closely adjacent stops are concentrated to one cabin which is promoted by a variable and controllable distribution bonus (Bvn). For this purpose, the estimated costs of lost time (GVK) for all passengers are calculated for each lift and for each storey call, these estimated costs of lost time (GVK) are reduced by a variable distribution bonus (Bv) concentrating adjacent stops to one cabin and a storey call is allocated for service by that lift which has the lowest reduced estimated costs of lost time (GVK red.min.). Both the estimated costs of lost time (GVK) and the variable distribution bonus (Bv) are calculated in accordance with closed formulae. To ensure that this method operates both with a high and with a low traffic load (Va), the variable distribution bonus (Bv) is corrected by means of a correction function (F[Va]) in accordance with the relation Bvn = Bv F[Va] in accordance with the traffic load (Va) as correction parameter (NFP) and thus the variable correctable distribution bonus (Bvn) defined (SR18). The correction function F[Va] can be determined by means of KI method according to Step 16 (SR16) or in accordance with expert programs according to Step 17 (SR17). <??>Keeping down the local accumulation of lift cabins is optimum for each traffic load (Va) with the traffic-dependence of the distribution bonus (Bvn). <??>This method can be applied to a large number of different service requests, allocation criteria and correction parameters (NFP). <IMAGE>

Description

The invention relates to a method and a device against local accumulation of elevator cars in an elevator group with variable traffic volume, in which the function of an elevator group by a suitable allocation of floor calls to elevators during call operation, with regard to one, by any combination and weighting of Elements from a predefined set of function requirements defined function profile is optimized and in which this suitable call allocation is determined and carried out based on an allocation criterion with regard to an allocation parameter with the same modifying bonuses and malusses according to a special strategy by an allocation algorithm, with a first function requirement being determined by the allocation criterion regarding the Allocation parameter is introduced into the allocation algorithm and at least one second function request by modifying the allocation parameter by means of a bonus s is taken into account to promote the corresponding functional feature or by means of a penalty to inhibit the corresponding complementary feature and the second functional requirement is to keep the local accumulation of booths low and is promoted by concentrating adjacent stops on a single booth by means of a variable distribution bonus.

With such methods, the allocation of the floor calls to the elevators for the call operation is selected such that the floor call operation takes place as a first function request, for example with a minimal estimated waiting time, and at the same time the local accumulation of cabins is kept low as a second function request. Depending on the amount of the distribution bonus, one of the two measures can preferably be weighted more.

Up to now, floor calls have been assigned to a group of elevator cars using a variety of strategies. A strategy has become known from US Pat. No. 4,790,412 which estimates the arrival time ETA of each elevator car for a specific floor call to be allocated. A value is calculated for each elevator car, which represents the time that the car in question is estimated to need to reach the calling floor in the correct direction of travel in order to serve the floor call. The floor call allocation is given to the cabin with the lowest ETA value. This strategy is based on calculating the estimated time of arrival ETA for each elevator for each floor call in the building and then assigning a specific floor call to the car with the lowest ETA.
The concept of cabin distribution in an elevator system is shown when the continuously changing types of elevator traffic are considered. If the sending strategy keeps the elevators well distributed for all conditions except for the morning "up" peak traffic in the building, they are in a better position to respond to future floor calls. In order to achieve this, certain group controls use procedures which "park" cabins in the building when the traffic has decreased. Such "parking" is however inefficient! The elevators do not do a useful job when they go to the parking floors, they just waste energy. The unnecessary wear and tear and the unnecessary wear on the cabins drive up the maintenance costs.
US Patent No. 4,790,412 shows a better way to minimize the local accumulation of elevator cars in an ETA dispatch strategy by introducing an algorithm to solve the distribution problem that is part of the actual allocation algorithm. The distribution algorithm improves the distribution by taking into account previous allocations of the cabins when new allocations are made.
In the allocation algorithm, the floors of the building are processed sequentially from bottom to top for "Up" floor calls and from top to bottom for "Down" floor calls. When calculating the ETA times for the cabins, assignments already made behind the floor currently being processed are taken into account.
This floor is called the "scan floor" and is a floor call floor on which the scanner has stopped to allocate or reallocate a floor call. This includes stops that a car has to make behind the scanning floor due to a car call. If a car is already obliged to stop behind the scanning floor, then this car is given a greater chance of receiving the allocation for the floor of the floor call currently being processed by calculating a dynamic additional value Tx and subtracting this additional value from the calculated ETA of the cabin. The same procedure is used when a car has an intermediate stop between the "current position" and the scanning floor. The "current position" or the selector floor (AVP floor) is the current floor position of the car when it is at rest or it is the floor on which the car can make a normal stop when it is in motion. This calculated additional dynamic value Tx favors the concentration of closely adjacent stops on one cabin and thus minimizes the local clustering of cabins. This dynamic additional value Tx can be said to be inversely proportional to a predefined travel distance of the elevator car. This predefined travel distance can be the same, regardless of whether the stopover results from a car call or an assigned floor call, by using the travel distance between the "current position" of the elevator car to be assigned and the scanning floor, as shown in the following formula I. .

It follows that the further the scanning floor is from the selector floor with regard to the driving distance to be covered from the cabin in order to fulfill its trip list, the smaller the amount of time that is deducted from the cabin's ETA. So if the car is a relatively long distance from the scanning floor, there is less chance of a build-up problem even if there is a stopover, and this is evidenced by the size of the additional value by making it insignificantly small.

As an alternative, the predefined travel distance can depend on whether the stopover results from an assigned floor call or from a cabin call. An assigned floor call can be reassigned again, especially if the cabin is a relatively large distance from the floor call floor. So if the stop has a floor call as the cause, the travel distance from the selector floor to the scanning floor is used. If, however, a car call is the cause of the intermediate stop, i.e. a stop that the car has to make, the additional value can be increased in order to favor the currently considered car with regard to the floor call allocation by making the predefined travel distance equal to the travel distance from the car call floor to the scanning floor. The use of a dynamic add-on value in accordance with U.S. Patent No. 4,790,412 prevents local cabin accumulations by concentrating closely adjacent stops on a particular cabin, which ensures better cabin distribution over the entire building without having to place false calls for parking floors. The use of this dynamic additional value increases the probability that the cabins are already adequately spaced from one another when they are stopped.

Although such a closely spaced landing call allocation concentrating on a single car is the local one Accumulation of elevator cars is significantly reduced, but there is a major disadvantage: As can be seen from equation (I), the size of the dynamic additional value Tx is not dependent on the traffic volume. Regardless of the number of calls in the system, the same additional value is calculated, ie the measure against the local accumulation of elevator cars is not tracked to the traffic volume. Although an allocation based on this strategy achieves good distributions when there is little traffic, it often leads to bad distributions of the cabins in the building when there is a lot of traffic. With increasing traffic, the cabins begin to accumulate locally, the redistribution of the cabins then depending on the randomness of the types of traffic and on the decrease in traffic volume. During times of high traffic, the result is poor service and an increase in the mean waiting time that a passenger has to wait for the service. The mean waiting time is an industry standard for measuring the efficiency of an elevator system.
The invention seeks to remedy this.

Accordingly, it is the object of the application according to the application to provide a method and a device for keeping the local accumulation of elevator cars equally small with each traffic volume in elevator groups and thereby reducing the mean waiting time. The method is intended to ensure a uniform elevator distribution, particularly in the case of high traffic volume, and to be designed as an integral part of the floor call allocation algorithm.
This object is achieved according to the invention with the means as characterized in the versions of the independent claims. Advantageous further developments are specified in the dependent claims.

Fig.1

ist eine vereinfachte, schematische Darstellung eines beispielhaften Aufzugssystems, welches die vorliegende Erfindung benützen kann.

Fig.2

ist ein vereinfachtes logisches Flussdiagramm eines Programmes, welches Stockwerkrufe einer Gruppe von Aufzügen A,B,C zuteilt.

Fig.3

stellt ein vereinfachtes logisches Flussdiagramm dar zur erfindungsgemässen Modifikation der Subroutine, um für eine Kabine die geschätzen Verlustzeitkosten (GVK) zur Bedienung eines bestimmten Stockwerkrufes zu berechnen.

Fig.4

zeigt ein Drei-Kabinen-Beispiel, um das grundlegende Konzept zu illustrieren.

The invention will be described using the example of a special GVK sending system. However, it should be understood that the invention can be used to improve any other type of delivery system. The invention will be better understood and features other than the foregoing will be understood when the following detailed description is read in conjunction with the figures which illustrate an exemplary embodiment of the invention:

Fig. 1

FIG. 10 is a simplified, schematic illustration of an exemplary elevator system that may use the present invention.

Fig. 2

is a simplified logic flow diagram of a program that allocates floor calls to a group of elevators A, B, C.

Fig. 3

represents a simplified logic flow diagram for the modification of the subroutine according to the invention in order to calculate the estimated lost time costs (GVK) for a cabin to service a particular floor call.

Fig. 4

shows a three-cabin example to illustrate the basic concept.

In order to illustrate an exemplary application of the present invention in detail, the disclosures in the above-mentioned Mac Donald US Pat. No. 4,790,412 dated December 13, 1988 and in the Swiss application 00 570 / 90-4 assigned to the same applicant with the title "Method and device for immediate assignment of destination calls to elevator groups, due to operating costs and variable bonus / malus factors "included as a reference in the description of the entire application.

In Fig. 1 , A, B, C denote the elevators of an elevator group, with one in each elevator Elevator shaft 1 guided car 2 is driven in a known manner by a drive 3 via a conveyor cable 4 and 16 floors E1 ..... E16 are served. The elevators may be of the hydraulic type or of the traction type as desired. Each drive 3 is controlled by a drive control known, for example, from European Patent No. 0 026 406, the setpoint generation, the control functions and the start-stop initiation being implemented by means of an industrial computer 5, and with 6 symbolizing the measuring and actuating elements are connected to the industrial computer 5 via a first interface IF1 and an elevator bus 7, which provides group control to the elevators A, B, C.

Each cabin 2 has a load measuring device 8, a device 9 signaling the respective operating state Z of the cabin, a stop indicator 10, and a car operating panel 11. The devices 8, 9, 10, 11 are connected to the industrial computer 5 via a cabin bus 12. Car calls are registered in the elevator cars with suitable push button arrangements designed as part of the car operating panel 11. They are then serialized and transmitted to the industrial computer 5 via the cabin bus 12 and interface CIF, together with other cabin-related data.
Floor calls are registered via suitable push buttons 13 located on the different floors E1 ... E16; an "up" floor call button 14 on the bottom floor E1, an "down" floor call button 15 on the top floor E16 and "up" and "down" floor call buttons 16 on each of the mezzanine floors E2 .. ..E15. Just like the car calls, the floor calls are serialized and then transmitted via floor bus 17 and input interface ICF to the industrial computer 5, where they use the distribution bonus Bv, which minimizes the local accumulation of elevator cars, in the sense of the required functional profile, to the individual Cabins 2 can be allocated for operation.

mit

B1...

: Bonusse

M1...

: Malusse

Derart berechnete Bedienungskosten KNR stellen ein Mass dar für die Bedienungsfähigkeit eines Aufzuges A,B,C... bezüglich eines Stockwerkrufes und hinsichtlich eines geforderten Funktionsprofiles einer Aufzugsgruppe. Ein Ruf wird dann demjenigen Aufzug A,B,C... zur Bedienung zugeteilt, der im Bedienungszeitpunkt voraussichtlich die grösste Bedienungsfähigkeit besitzen wird, d.h. dessen Zuteilungsparameter ZTP dem Zuteilungskriterium ZTK voraussichtlich am besten entsprechen wird und der somit die geringsten Bedienungskosten KBN aufweisen wird.2 shows the structure and the sequential sequence of the floor call allocation algorithm SZA with its two subordinate algorithms for the bonus tracking NFA and the loss time cost calculation KBA and the subroutine NVa for tracking the distribution bonus Bv according to the traffic volume Va.
It can initially be assumed that the industrial computer 5 is informed about the operating status of the elevator groups A, B, C via the cabin bus 12, elevator bus 7 and floor bus 17. He therefore knows the load, the position and the operating status of the car 2, the operating status of the drive 3 at every point in time for each elevator A, B, C and he also has information about the current traffic situation and the currently valid bonuses / penalties B1. ../ M1 .... On the basis of this information, the floor allocation algorithm SZA is able to allocate newly entered floor calls to the elevators A, B, C in the sense of predetermined criteria, ie to determine a call allocation that is optimal according to these criteria. These criteria are essentially functional requirements FA1, FA2, ... for the function of the elevator group. Such a call allocation takes place continuously at the processing speed of the industrial computer 5 as part of the sequential call processing of all floors E1, E2, .., from the first scanning of the corresponding floor call until the point in time immediately before it is operated.
The basis of the call allocation are defined by an allocation parameter ZTP - which can be modified - operating costs KNR which are calculated as follows:

$\text{KNR = ZTP possibly modified by (B1 ... / M1 ...) (II)}$

With

B1 ...

: Bonuses

M1 ...

: Malusse

Operating costs KNR calculated in this way represent a measure of the operability of an elevator A, B, C ... with regard to a floor call and with regard to a required functional profile of an elevator group. A call is then allocated to the elevator A, B, C ... for operation, which is likely to have the greatest operability at the time of operation, i.e. whose allocation parameter ZTP is expected to best match the allocation criterion ZTK and which will therefore have the lowest operating costs KBN.

The preferred embodiment variant chosen to represent the method according to the invention against the accumulation of elevator cars will now be explained using the floor call allocation algorithm SZA according to FIG. 2. This preferred embodiment variant is characterized in that the estimated loss time costs GVK, known as operating costs KNR, are equal to the sum of the estimated loss times of all passengers GVK expressed in passenger seconds, and that, to keep the accumulation of elevator cars low, a variable, that is, according to a tracking function F (Va) des Traffic volume Va trackable distribution bonus Bvn is provided, which is calculated according to a special formula and which multiplies the estimated loss time costs GVK to modified estimated loss time costs GVKmod. reduced.
In the preferred embodiment variant, this results in the following formulas for the modified, estimated cost of loss GVKmod. and the distribution bonus Bvn:

$\text{GVKmod. = GVK Bvn (III)}$

wobei K eine geeignet gewählte Konstante darstellt.According to FIG. 2, the allocation process begins with a first step SR1, which consists of scanning a registered floor call that has not yet been served. This floor call is now not assigned arbitrarily, but in the sense of the two functional requirements FA1 and FA2 which are the basis of the group function. For this purpose, the functional requirements FA1, FA2 ... are hierarchically arranged in a second step SR2 and divided into two groups: the first group for priority function requirements contains FA1, the second group for subordinate function requirements contains FA2. This division is necessary because in the cost calculation described below after steps SR8 and SR9, a distinction is made between these two groups, in that the priority function request FA1 is represented by the estimated loss time costs GVK and the subordinate function request FA2 is represented by a trackable distribution bonus Bvn that acts on GVK . Step 3 SR3 is about determining whether the prerequisites for using the method according to the invention are present. This method consists in keeping the local accumulation of booths low by concentrating closely adjacent stops on the same booth. However, if there are no stops between the STW.a scanning floor and the Stw.s selector floor, there is no reason to use this procedure. In this case, the floor call allocation is based on the unmodified estimated loss time costs GVK. These are calculated in step 8 SR8 according to the special cost formula as set out in the applicant's European Patent No. 0 032 213 on pages 4 and 5 and the floor call - after step 10- the elevator with the lowest unmodified estimated loss time costs GVKmin. assigned for operation.
However, if SR3 is found in step 3 between the scanning floor Stw.a and the selector floor Stw.s Halt, the prerequisites for using the method according to the invention are given. Therefore, in a next step 4 SR4, the minimization of the accumulation of elevator cars, ie their Even distribution in the elevator shaft ensures distribution bonus Bv calculated according to the following formula:

where K represents a suitably chosen constant.

Bvn

: nach dem Verkehrsaufkommen Va nachgeführter, variabler Verteilbonus Bv.

As an essential feature of the innovation according to the invention, the variable distribution bonus Bv can now also be tracked, ie it is tracked according to a tracking parameter. According to step 5 SR5, this can relate to an individual elevator or to the entire elevator group. In the present example, the current traffic volume Va is a group-related tracking parameter, and the corresponding tracking function F (Va) is determined after step 7 SR7 using a special subroutine, which is shown in more detail in FIG. Regardless of whether it is an elevator-related or a group-related tracking parameter, step 9 SR9 in turn leads to the entry into the cost calculation algorithm KBA. In this second case, ie if a variable and trackable distribution bonus Bvn is available, the modified estimated loss-time costs GVKmod are calculated for all elevators A, B, C ... according to the following formula:

$\text{GVKmod. = [GVK (inside) + GVK (outside)] Bvn (VI)}$

Bvn

: variable distribution bonus Bv adjusted according to traffic volume Va.

Finally, in a last step 11 SR11, analogously to step 10, the present floor call is allocated to that elevator of elevator group A, B, C, .... the lowest modified, estimated loss time costs GVKmod.min. having.

FIG. 3 explains the subroutine for adaptive tracking of the distribution bonus Bv as shown in step 7 SR7 of FIG. According to step 5 SR5, the group-related traffic volume Va is again determined as the tracking parameter. Next, in step 14 SR14, the tracking of the distribution bonus Bv according to the traffic volume Va is represented in a formal manner. The task of the subroutine is now to determine the tracking function F (Va) in step 14 SR14, according to which the traffic volume Va tracks the distribution bonus Bv. For this purpose, a distinction is made in step 15 SR15, namely derivation via artificial intelligence AI or use of existing expert knowledge. The function F (Va) is therefore optionally determined in step 16 SR16 using AI methods and in step 17 SR17 using expert programs. In both cases, a function F (Va) results with which the variable distribution bonus Bv can be tracked according to the traffic volume Va. F (Va) is the preferred embodiment variant e.g. monotonically increasing function of Va. This is represented by the formula in step 18. According to this, the tracking distribution bonus Bvn is obtained by multiplying the tracking function F (Va) on the variable accounting bonus Bv calculated according to formula V. According to formula VI, this affects the modification of the estimated loss time costs: the higher the traffic volume, the greater the modification of the GVK of the corresponding cabin in subsequent, neighboring allocations.

To illustrate the concept even further, consider the three-cabin example according to FIG. 4. The "up" arrows 18 and the "down" arrows 19 represent floor calls. It should be noted that these arrows 18, 19 represent simple floor calls in the case of low traffic volume and multiple floor calls in the case of high traffic volume.
FIGS. 4a and 4b represent allocations in the case of low and high traffic volume which were made with an allocation algorithm which only used a variable but not a traceable distribution bonus Bv. In In this example, Fig. 4a represents a desirable distribution with little traffic. Fig. 4b clearly shows the deterioration. By taking the traffic volume according to equation IV into account, the variability of the distribution bonus Bvn according to the invention pushes the distribution back to the desired one according to FIG. 4 a and thus ensures better distributions with each traffic volume.

Claims (7)

Method and device against local accumulation of elevator cars in an elevator group with variable traffic volume, in which
the function of an elevator group
is optimized by a suitable allocation of floor calls to elevators during call operation with regard to a function profile defined by any combination and weighting of elements from a given set of function requests (FA1, FA2 ...) and in which this suitable call allocation is based on an allocation criterion (ZTK ) with respect to an allocation parameter (ZTP) with the same modifying bonuses / malusses (B / M) according to a special strategy by an allocation algorithm (ZTA) is determined and carried out, whereby a first function request (FA1) by the allocation criterion (ZTK) with respect to the allocation parameter (ZTP ) is introduced into the allocation algorithm (ZTA) and at least one second function request (FA2) by modifying the allocation parameter (ZTP) by means of a bonus B to promote the corresponding functional feature or by means of a malus M to inhibit the corresponding complementary feature is taken into account and wherein the second function requirement (FA2) consists in keeping down and the local bunching of cages adjacent concentration by holding a single car by means of a variable Verteilbonus Bv is conveyed,
characterized, - that the functional requirements (FA1, FA2, ...) defining the functional profile of an elevator group are hierarchically arranged and are divided into at least two groups: primary functional requirements and subordinate functional requirements, - That as a priority function request FA1, the floor calls with minimal estimated Loss-time costs of all participating road users (GVKmin) are served whereby the estimated loss-time costs (GVK) of each lift serve as allocation parameters (ZTP) and the allocation criterion (ZTK) consists in minimizing the estimated loss-time costs (GVK) assigned to a call service . - that as a subordinate functional requirement FA2,
the low level of local accumulation of cabins is ensured by allocating closely adjacent floor calls to the same cabin for operation, for which purpose the allocation algorithm (ZTA) provides for the estimated loss time cost (GVK) of the elevator concerned to reduce the variable distribution bonus Bv. - that the variable distribution bonus Bv for minimizing the local accumulation of cabins in their numerical value in groups, the traffic volume (Va)
the elevator group, is adaptively tracked. - That the variable and trackable distribution bonus Bvn acts as a subtrahend or as a multiplier on the estimated loss time costs (GVK) and reduces them subtractively or multiplicatively. The method of claim 1
characterized,
that the priority function request FA1 is to serve the floor calls with, for example, a minimum estimated waiting time (GWZmin), the estimated waiting time (GWZ) serving as the allocation parameter (ZTP) for each individual elevator and the allocation criterion (ZTK) in minimizing the estimated waiting time assigned to a call service ( GWZ) exists. The method of claim 1
characterized,
that the operating requirements that are assigned to the individual elevators in terms of a required group function for operation are destination calls. The method of claim 1
characterized,
that the numerical value of the variable and trackable distribution bonus Bvn is indirectly proportional to the distance, i.e. to the number of floors between the scanning floor (Stw.A) and the selector floor (Stw.S), and directly proportional to a function F of the traffic volume Va of the elevator group and the traffic volume Va in groups is updated as follows:

The method of claim 1
characterized,
that the numerical value of the variable distribution bonus Bv is tracked in groups according to a parameter of the elevator group or in part according to a parameter of an individual elevator. Method according to claim 5
characterized,
that artificial intelligence methods or expert programs are used to track the variable distribution bonus Bv in groups and by elevator. The method of claim 1
characterized,
that the assignment of an operating request to an elevator, which guarantees the required functional behavior of an elevator group, is determined only once, immediately after registration of the operating request, or is continuously redetermined until immediately before it is operated by an elevator.

Images