WO2009078834A1

WO2009078834A1 - Elevator system traffic profile generator

Info

Publication number: WO2009078834A1
Application number: PCT/US2007/025743
Authority: WO
Inventors: Robert E. Labarre; Theresa M. Christiy
Original assignee: Otis Elevator Company
Priority date: 2007-12-17
Filing date: 2007-12-17
Publication date: 2009-06-25

Abstract

A method for method for determining a traffic list for elevator cars operating in an elevator system in a building in which system selected performance events associated with those elevator cars and the corresponding times of occurrence thereof can be recorded in a data storage facility during operations of those elevator cars over selected time durations based on acquiring from the data storage facility a chronologically ordered list of operating elevator car performance events and the corresponding times of occurrence thereof and then forming an initial passenger list based on assuming there is a single passenger in each of the operating elevator cars departing from origin building floors after arriving there in response to each summons made in the system from those floors for transportation service to other building floor destinations. The number of passengers in each of the operating elevator cars following such departures is increased by some requests made in each such elevator car for transportation service to a selected building floor destination and by forming chance based additions through using numerical weightings operating similarly to probability distributions to allow probabilistically selecting to include additional passengers and their destinations.

Description

ELEVATOR SYSTEM TRAFFIC PROFILE GENERATOR

BACKGROUND

The present invention relates to elevator arrangements for buildings and, more particularly, to methods for developing approximate passenger, or "traffic", lists associated with use of such arrangements from incomplete data characterizing such use.

Transferring people from one location to another in buildings of any significant size with multiple floors therein is primarily accomplished through the use of elevators provided in such buildings. Such transfers of people have the elevators of a building transporting passengers therein from one building location to another, and patterns of such transfers result which are peculiar to the building, its elevators, its temporary and longer term inhabitants along with their corresponding building locations, the various times of the day during the differing days of the week, etc. Knowledge of, or good estimates of, such people relocation, or traffic, patterns is important to designers of new or replacement buildings, or designers of modifications to existing buildings, and to managers of buildings in determining how many elevators to provide in a new or remodeled building or to have in operation at any given time in such a building. Traffic patterns are often described as fractions of the number of longer term building inhabitants present at some location in succeeding five minute intervals over some duration, and smaller fractions at such intervals in the building lobby versus larger fractions, for instance, indicates relatively fewer elevators are needed at that lobby.

Traffic lists are one kind of representation of such traffic patterns with each being essentially a list of each passenger that uses (or is expected to use) the elevator system of a building (or proposed building) of interest. In a complete traffic list, each passenger is essentially described by the time that he or she arrives in a floor hallway, the floor from which service is requested (origin) and the floor to which service is requested (destination). However, getting accurate traffic lists from existing buildings is difficult because sensors sufficient to provide complete information are generally not provided in the great majority of existing conventional building elevator systems. Such conventional elevator systems typically have a computer linked to the elevator system controller that can record various hall call button and car call button pressings, car positions and car load weights, and the times of occurrences thereof.

So, although conventional elevator systems, for example, can estimate when the first person in a floor hallway arrives (because that person presses one of the up or down hall call buttons at the elevator location in that hallway) the system cannot determine whether one, two, three or more people arrive afterwards wanting to go in the same direction and, therefore, have no need to press such a hall call button.

Similarly, such a conventional elevator system cannot necessarily determine destinations of elevator car passengers. Assuming an elevator car is simply empty when it arrives, then the car call button that is pressed inside the car by a passenger will inform the system of the destination to which at least one car passenger wants to travel. If, however, the elevator car arrives at a floor hallway with several passengers already in it, and has, for example, car call buttons for floors 3, 4 and 5 are already pressed to select those floors for car stops, then, if a new passenger boards the elevator but does not press a car call button, the system cannot determine if that new passenger wants to go to floor 3, 4 or 5.

As a' result, the data collected from operation of conventional elevator systems can, based only on the information therein, be used to provide, at best, just incomplete traffic lists for the system. Thus, there is a desire for a methodology to further develop such otherwise incomplete traffic lists into a more suitable list.

SUMMARY The present invention provides a method for determining a traffic list for elevator cars operating in an elevator system in a building having a tabulation of passengers chronologically listed therein with arrival times thereof at corresponding floors of origin for transport to floors of destination selected by them in those operating elevator cars and in which system selected performance events associated with those elevator cars and the corresponding times of occurrence thereof can be recorded in a data storage facility during operations of those elevator cars over selected time durations based on acquiring from the data storage facility a chronologically ordered list of operating elevator car performance events and the corresponding times of occurrence thereof. This is followed by forming an initial passenger list based on assuming there is a single passenger in each of the operating elevator cars departing from origin building floors after arriving there in response to each summons made in the system from those floors for transportation service to other building floor destinations, and increasing the number of passengers in each of the operating elevator cars following such departures thereof by that number of any requests made in each such elevator car, beyond one such request, for transportation service to a selected building floor destination prior to those elevator cars subsequently arriving at a next building floor. The possibility of there being additional passengers boarding at elevator car floor stops is included by forming a chance based traffic list through using numerical weightings operating similarly to probability distributions to allow probabilistically selecting to include additional passengers and their destinations.

BRIEF DESCRIPTION OF THE DRAWINGS Figures IA, IB and 1C show a flow chart displaying the method of the present invention for developing traffic lists.

DETAILED DESCRIPTION

There is described in the following the forming of an incomplete traffic list from the data collected in operating a conventional elevator system, and developing that list into a full traffic list for that system based to at least a significant degree on chance, i.e. based to a significant degree on a more or less unpredictable outcome of a selection made in an uncertain process. The incomplete list is based on the conventional elevator system having, or at least being conducive to temporarily having added thereto, a data collection facility capable of providing an events record of (a) the hall calls initiated by floor or hall occupants, i.e. potential passengers or just passengers, arriving on corresponding ones of the numbered floors through pushing the hall call button for requesting elevator car service for that floor along with the time of that button push, (b) the times that each of the numbered elevator cars arrives at one of the numbered floors, (c) the car calls initiated by passengers in the various numbered elevator cars through pushing a car call button in the corresponding car to select a destination one of the numbered floors along with the time of that button push, (d) the times that each of the numbered elevator cars leaves one of the numbered floors, and (e) the one of a set of distinct car weight categories that the weight of each of the numbered elevator cars is in as that car leaves one of the numbered floors. The uncertain process trials that are used to develop the incomplete traffic list into a chance based, or probabilistic trial based, full traffic list are described as being provided on the basis of a computer based random, or pseudorandom, number generator but could instead be accomplished using well known Monte Carlo techniques.

There is shown immediately following, a Table 1, that has tabulated therein, as an example of an elevator operation scenario, a succession of such recordable events that are to be taken as having been recorded by the data collection facility in a conventional elevator system. These recorded events are taken to be the result of the indicated action occurrences having happened in an example building elevator system during a limited duration sequence of operations of that elevator system. There are two elevator cars in this example building and both of them are parked at the building lobby as the example scenario begins. This scenario will be used to illustrate the method of the present invention for generating a probability based passenger, or traffic, list but the listed events are not exhaustive of the kinds of event combinations that can occur in actual practice.

Table 1 : Recorded Events in Example Scenario

passengers on board (Load weight value is 2 since there are 4 passengers on board, but it is not reported since it has not changed since last report.)

00:31 Car 2 arrives at floor 7 and deboards a passenger

00:32 Car 1 arrives at floor 7 and B,D,G deboards passengers

00:41 Car 2 leaves floor 7 with passenger on board (Load weight is 0 since there is only 1 passenger on board)

00:42 Car 1 leaves floor 7 with passenger on board (Load weight is 0 since there is only 1 passenger on board.)

00:43 Car 1 gets to floor 8 and deboards passenger

00:46 Car 2 gets to floor 12 and deboards passenger

00:53 Car 1 leaves floor 8 and heads to floor 5 to pick up passengers (Load weight is 0 since there are no passengers on board, but it is not newly reported since it has not changed.)

00:56 Car 1 gets to floor 5 and boards H,I,J,K passengers

01:06 Car 1 leaves floor 5 with H,I,J,K passengers on board (Load weight is 2 since there are 4 passengers on board.)

1:11 Car 1 gets to the lobby and H,I,J,K deboards passengers.

Table 2 shows an exact passenger or traffic list that is in accord with the example scenario described in Table 1 that could be obtained if there were available a data collection system capable of recording every action of every passenger with respect to that passenger's encounter with the elevator system rather than having to rely on the limited data collection system described above. A typical passenger list includes a unique identifier for each passenger (usually assigned to each passenger as 1,2,3... to represent the first, second, third, etc. passenger). In Table 2, the passengers are each identified with corresponding one of A, B, C to make them easier to identify in the presence of so many other uses there of numbers therein. In addition to the unique passenger identifier, each passenger record includes the time the passenger arrived in the hallway of a floor to await an elevator car, the floor of origin (i.e. where the passenger waits for the elevator), and the floor of destination (i.e. where the passenger wants to get out of the elevator car taken to that floor).

Table 2: Exact Passenger List for Example Scenario

The following assumptions are used for the example scenario: (1) There are three distinct load weight categories and 0 or 1 passenger in an elevator car leads to the car and passenger weight load being in the smallest load weight category (Category 0),

(2) 2 or 3 passengers on an elevator car leads to the car and passenger weight load being in the next and second smallest load weight category (Category 1), (3) 4 or more passengers on an elevator car leads to the car and passenger weight load being in the next and third smallest load weight category (Category

2),

(4) each elevator car takes 10 seconds for every stop,

(5) each elevator car takes 1 second to travel through a floor, (6) a car at the lobby will wait 10 seconds after receiving a destination floor assignment at the lobby before it leaves the lobby,

(7) the relevant load weight category is recorded when a car leaves the floor after stopping, but only if the load weight has changed since the last time the car left a floor after stopping, (8) typical conventional elevator system Door Open and Door Close messages are eliminated for brevity,

(9) elevator car position messages are condensed for clarity, and

(10) each elevator car can hold a maximum of eight passengers.

Table 3 is a table that shows the data that will be in the conventional elevator data collection system database if the example scenario of Table 1 occurs. Table 3 is organized chronologically by times of occurrences of the various recorded actions.

Table 3: Data Collection System Database Resulting from Data Recorded During Example Scenario

Figures IA, IB and 1C show a flow chart, 10, displaying the method of the present invention for developing traffic lists on a probabilistic trials basis in a traffic list developmental computer system from the information provided through the recordations occurring in the data collection facility in a conventional elevator system as assembled in the database therein, starting in a start balloon, 11, in Figure IA. This flow chart will be described below in connection with the example scenario of Tables 1 and 3 resulting in a probabilistic based traffic list that will be compared with the exact traffic list of Table 2 following the completion of this developed traffic list in the description below.

The probabilistic based traffic list development begins with the list developmental computer system obtaining the data previously stored in some data collection facility database concerning the relevant parameters of the elevator system of interest in the building of interest such as car size, maximum loading and load weight settings from the sample building personnel. This is undertaken in a decision diamond, 12, determining if such data is already in the list developmental computer system or must be retrieved from a database facility, 13, shown in this example also serving as the conventional elevator system data collection facility installed in the example building of interest, from which this retrieval is made.

In addition, the list developmental computer system must obtain the recorded events data stored in the conventional elevator system data collection facility installed in the example building of interest for a time duration matching the time duration over which the probabilistic based traffic list is selected to be developed. This is undertaken in a decision diamond, 14, determining if such data is already in the list developmental computer system or must be retrieved from database facility 13, from which this retrieval is made which, here, is the data listed in Table 3 for the example scenario.

The record entries in the conventional elevator system data collection facility database for each elevator car in the elevator system of interest during the time duration of interest will be evaluated separately and then merged at the end of the process to obtain the complete probability based traffic, or passenger, list for that time duration. Thus, initially, the number N of elevator cars in operation in the elevator system of interest is entered into the list developmental computer system with those cars being identified by the number assigned to them in the data collection facility database assumed in flow chart 10 to be cars 1 through N, as indicated in a performance block, 15, in Figure IA, and with N = 2 in the example scenario.

For elevator car CAR_k, with that being CAR] in the first instance, the list developmental computer system sorts through the entries in the recorded events data for the record entries for CAR^ and compiles a list of all of the hall call assignment recorded events therefor to form a hall call list as indicated in another performance block, 16, in Figure IA. Each record entry in this latter list for CAR_k will contain a field having the time that the hall call therein was assigned thereto, a field having the direction of elevator car travel indicated by that hall call, and a field having the floor hall location at which that hall call originated. Thereafter, the list developmental computer system sorts this hall call list by assignment times and forms a hall call assignment list with the record entries therein being in chronological order by assignment times as indicated in a further performance block, 17, in Figure IA. Table 4 shows the resulting hall call assignment lists resulting from these operations on the data collection system database of Table 3 for the example scenario. As the foregoing data manipulations will be repeated for each operating elevator car in the elevator system of interest, the results are shown in Table 4 for both of the cars in this example scenario. Table 4: Car Hall Call Assignments Lists for Example Scenario For Car 1

For elevator car CAR_k, with that being CAR₁In the first instance, the list developmental computer system sorts through the entries in the recorded events data for the record entries for CARk and compiles a list of all of the car call recorded events therefor to form a car call list as indicated in a subsequent performance block, 18, in Figure IA. Each record entry in this latter list will contain a field having the time that each car call was entered by a passenger in CAR_k and a field having the destination indicated by that car call. Thereafter, the list developmental computer system sorts this car call list by call times and forms a car call initiation list with the record entries therein being in chronological order by call times as indicated in a further performance block, 19, in Figure IA. Table 5 shows the resulting car call initiation lists resulting from these operations on the data collection system database of Table 3 for the example scenario. As the foregoing data manipulations will be repeated for each operating elevator car in the elevator system of interest, the results are shown in Table 5 for both of the cars in this example scenario. Table 5: Car Call Initiations Lists for Example Scenario For Car 1

The list developmental computer system then forms a merged list as an initial passenger list for CAR_k in a following performance block, 20, using the record entries from the hall call assignment and car call initiation lists previously compiled for CAR_k that are ordered chronologically therein by the assignment and call times, respectively. This merged list for each record entry therein will (a) contain a field having a passenger arrival time which is the assignment time from the hall call assignment list for entries corresponding to that list, (b) contain a field having an origin which is the floor location from which the hall call was made for entries corresponding to the hall call assignment list, and (c) contain a field having a destination which is an arbitrarily chosen placeholder symbol for the unknown destination for entries corresponding to the hall call assignment list, and (d) contain a field having a passenger arrival time which is an arbitrarily chosen placeholder symbol for entries corresponding to the car call initiation list, (e) contain a field having an origin which is an arbitrarily chosen placeholder symbol for entries corresponding to the car call initiation list, and (f) contain a field having a destination which is the floor destination for entries corresponding to the car call initiation list. Table 6 shows the resulting merged lists resulting from these operations on the lists of Tables 4 and 5 for the example scenario. The number "255" has been chosen as the arbitrary placeholder in each instance the need for a placeholder arises. As the foregoing data manipulations will be repeated for each operating elevator car in the elevator system of interest, the results are shown in Table 6 for both of the cars in this example scenario. Table 6: Merged Lists as Initial Passenger Lists for Example Scenario Car l

For elevator car CAR_k, with that being CARjin the first instance, the list developmental computer system sorts through the entries in the recorded events data for the record entries for CAR_k and compiles a list of all of the car floor stop events therefor to form a car halts list as indicated in a subsequent performance block, 21, in Figure IB. The transition path from Figure IA to Figure IB is indicated by a transition balloon, A, in each figure. Each record entry in this latter list will a field having the floor at which CAR_k stopped, a field having the reason for CAR_k stopping at that floor, a field having the time at which CAR_k arrived at that floor, a field having the time CAR_k left that floor, a field having a listing of the car calls still pending for service and any car calls subsequently initiated prior to the next stop made by CAR_k, and the load weight category code associated with CARk at its leaving from the floor. The list development computer system must compile and coordinate different record entries in the recorded events data for each car stop based on the times associated with the record entries and determine which data is relevant to each car stop. Load weight category code values are only reported when they change, so, if there is no load weight category code associated with a car leaving a floor, then the most recent load weight category code is retained. Thereafter, the list developmental computer system sorts this car halts list by car floor stop times, i.e. the car floor arrival times, and forms a car stop list with the record entries therein being in chronological order by stop times as indicated in a further performance block, 22, in Figure IB. Table 7 shows the resulting car stop lists resulting from these operations on the data collection system database of Table 3 for the example scenario. The entries "NA" mean nonapplicable as the cars have completed the services requested during the time duration of this example. As the foregoing data manipulations will be repeated for each operating elevator car in the elevator system of interest, the results are shown in Table 7 for both of the cars in this example scenario. Table 7: Car Stop Lists for Example Scenario Car l

Car 2

The stop list determined in performance block 22 is then used by the list developmental computer system in a further performance block, 23, to supply the information to replace some of the unknown values present in the in the initial passenger list formed by the merged list determined in performance block 20 to form an updated passenger list based on assuming there is a single passenger associated with each car call. The list developmental computer system associates each car call in the merged list of performance block 20 with the hall call in that merged list that is from the floor location at which CAR_k was most recently stopped prior to the time of that car call as determined from the stop list of performance block 22. The updated traffic or passenger list, i.e. the single passenger per hall call supposition passenger list, is thus formed with record entries for each car call and its so associated hall call plus the hall call entries for which there is no associated car call in hall call chronological order. (If a hall call is answered without generating a new car call to be associated therewith, then the passenger associated with the hall call must be going to one of the existing car call destinations.) In the situation in which there are two hall calls each entered from the same floor and each assigned to CAR_k but which have not yet been responded to, determine the direction of car travel from the car calls on the stop list and associate those car calls for one direction with the one of those two hall calls requiring that direction. Each hall call entry in this list, and any car associated therewith, will (a) contain a field having a passenger arrival time which is the passenger arrival time for that hall call in the merged list, (b) contain a field having an origin which is the origin for that hall call in the merged list, and (c) contain a field having a destination which is the destination for that car call in the merged list or, if there is no car call associated with the hall call entry, then the pending destinations from the stop list when the car was at the origin for this hall call entry. The passenger arrival times for each passenger associated with a hall call is taken in the foregoing to be the passenger arrival time from the merged list but could instead be distributed (perhaps probabilistically) over the time between the initial hall call time at the floor location and the time at which CAR_k arrives that floor location to thereby vary the passenger waiting time system performance parameter at that location.

As seen in the foregoing, both hall call and car call information is needed to construct a complete record for an individual passenger. The hall call information provides the origin (i.e., the starting floor for that passenger) and the car call information provides the destination (i.e., the exit floor selected by that passenger). Thus, in the example scenario, pairs of entries in the merged list in Table 6 that formed the initial passenger list are combined to represent one complete passenger record in the subsequently developed single passenger per hall call supposition passenger list of performance block 23 formed in a corresponding table, Table 8. Hence, in a first example in the example scenario, the first two records in Table 6 are combined to make one complete record as the first entry in the corresponding single passenger per hall call supposition passenger list of Table 8 below. Again, the foregoing data manipulations will be repeated for each operating elevator car in the elevator system of interest, and so the results are shown in Table 8 for both of the cars in this example scenario.

In addition, however, the stop list of Table 7 is needed to correctly associate a car call with its proper hall call in the example scenario. Consider the following example provided separately from the example scenario:

(a) elevator car 1 is stopped at floor 1,

(b) an up direction hall call is entered at time = 00:05 at floor 2 and assigned to car 1 (a more complete data collection system would know that the person at floor 2 wants to go to floor 7, but the conventional system does not yet know that), and

(c) an up direction hall call is entered at floor 3 at time = 00:07 and is assigned to car 1 (a more complete data collection system would know that there are two people at floor 3 who want to go to floors 5 and 6, respectively, but the conventional system does not yet know that). This example data leads to the resulting corresponding hall call assignment list (like that shown in Table 4 for the example scenario) then appearing as:

00:05 up 2 255 00:07 up 3 255.

Also, this further leads to a resulting corresponding car call initiation list (like that shown in Table 5 for the example scenario) that is in accord with the hall call times and the subsequent car floor arrival times can be chosen consistently with the foregoing as:

00:15 7

00:23 5 00:26 6.

The hall call and car call information in the hall call assignment list and the car call initiation list in the separate example just set forth above, used alone, leaves unclear just which car calls are to be associated with which hall calls. Adding stop information, however, allows the proper association using the procedure described above of associating a car call with the hall call at the floor at which the elevator car was most recently stopped prior to the initiation of the car call. Thus, if car 1 stopped at floor 2 at 00:12 and stopped at floor 3 at 00:22, as is consistent with the hall call assignment list and the car call initiation list set forth above, then the car call for floor 7 should be associated with the hall call at floor 2 and the other car calls should be associated with the hall call at floor 3 (two passengers on at floor 3). If, instead, car 1 stopped at floor 2 at 00: 14 and at floor 3 at 00:25, as is also consistent with the hall call assignment list and the car call initiation list set forth above, then the car calls for floors 7 and 5 should be associated with the hall call at floor 2 (two passengers on at floor 2) and the remaining car call for floor 6 should be associated with the hall call at floor 3.

Returning to the example scenario, if a hall call is satisfied by the arrival at the floor from which the hall call was made but without there being a new car call subsequently made, presumably the passenger entering the car at that floor must be going to one of the pending car call destinations. This is the situation represented by row 4 in the single passenger per hall call supposition passenger list of Table 8 for car 1. There was no new car call initiated when car 1 was stopped for the up direction hall call at floor 5. So, the entering passenger at floor 5 is assumed to intend to go to one of the previous car call destination floors there pending because of a corresponding previously made car call by an earlier boarding passenger. The car stop list of Table 7 shows that when car 1 was stopped at floor 5 there were pending car call destinations for floors 7 and 8. The assumption that the passenger entering at floor 5 intends to go to either floor 7 or floor 8 results in those floors being listed in the destination field for the car 1 hall call time entry of 00: 18 in the single passenger per hall call supposition passenger list of Table 8 for car 1. The further determination of whether that passenger shall be taken as traveling to floor 7 or floor 8 will be described below. Table 8: Single Passenger Per Hall Call Supposition Passenger Lists for Sample Scenario Car 1

Thus, the single passenger per hall call supposition passenger lists of Table 8 represent a rudimentary traffic, or passenger, list constructed for each elevator car in the example scenario in accord with performance block 23 of Figure IB. However, since the data collection facility in a conventional elevator system cannot detect the number of people associated with a specific hall call or car call, there remains the possibility that there are more passengers encountering the elevator system of interest during the time duration over which the single passenger per hall call supposition passenger list has been developed for CAR_k than those that are represented in that list. The information collected by the conventional elevator system data collection facility can, however, be used with added passenger numbers numerical weighting distributions behaving like assumed probability distributions as a basis to formulate a additionally developed traffic, or passenger, list for each operating car. Such a list is likely to be more representative of the elevator system passenger encounters that took place over the time duration of the collected information being used in developing such a list than is the single passenger per hall call supposition passenger list above developed for each operating car alone over that duration.

This additionally developed traffic, or passenger, list is provided through the list developmental computer system determining whether any of the car call and associated hall call combinations in the single passenger per hall call supposition passenger list for CARk found in performance block 23 could possibly represent also one or more additional passengers than the one passenger accounted for there, and then using such an assumed weight (probability) distribution as a basis for (a) selecting the number of additional passengers to be added to a further developed passenger list out of the number those additional passengers possible and for (b) selecting destinations for such added passengers out of those destinations therefor possible. The number of additional passengers that could be on CAR_k after a floor stop can be estimated by the list developmental computer system, as indicated in a further performance block, 24, in Figure IB, through subtracting from the known physical passenger capacity of that car the number of pending and new car calls as that car leaves that floor that are set out in the single passenger per hall call supposition passenger list for that car, and then choosing the maximum number in that difference that is consistent with the weight load category for that car leaving that floor from the car stop list therefor. This result can then be checked for consistency with the recorded events for the next floor stop for CAR_k and adjusted accordingly if need be.

Thereafter, each of the possible added passenger numbers from one to the maximum number possible found by the list developmental computer system for CAR_k is taken to be a passenger event and a corresponding numerical weight (probability) for that passenger number, or event, is judged and assigned thereto based on an assumed event weight (probability) distribution such that the sum of the weights (probabilities) for all such events equals 1. A random number generator in the list developmental computer system that can provide a sequence of numbers more or less randomly, but limited to some numerical generator range length, is then selected to be used. Each possible passenger event for CAR_k at the floor stop is assigned to a corresponding one of a sequence of subranges of numbers in the random number generator range with each being of a relative numerical length corresponding to the magnitude of the associated event probability, as further indicated in block 24, so that the sequence of these subranges together matches the random number generator range length. This random number generator is then operated by the list developmental computer system to more or less randomly issue a number which will thus be in one of the passenger event subranges to thereby select that passenger event and so select the number of additional passengers to be added to the further developed traffic list as indicated in a following performance block, 25.

In the example scenario, the car stop list of Table 7 shows car 1 initially leaving the lobby (or floor 0) having a load weight code indicating that there are 4 or more people on board that car. Three of those people are accounted for in the single passenger per hall call supposition passenger list of Table 8 for car 1 because there is one person associated there with each of the three corresponding car calls listed there. Assuming that the elevator car maximum capacity is 8 passengers, along with the load weight indication that there are at least four people on board, allows concluding that five possibilities for additional passengers being on board exist. That is, there are the possibilities of one person, two people, three people, four people or five people being on board car 1 in addition to the three already accounted for. The car stop list of Table 7 also shows that when car 1 leaves floor 5 it still has 4 or more people on board. On the assumption that one passenger gets on at floor 5 and another passenger gets off at floor 5 then the passenger numbers in car 1 before and after stopping at floor 5 are consistent. Thus, the load weight categories for subsequent stops have an effect in determining how many additional passengers might be added to the system and must be checked to assure car passenger numbers consistency. Assuming the passenger event probabilities to be equal, i.e. a discrete uniform probability distribution, (although any discrete probability distribution can be used, as desired) each of the five passenger event possibilities has an equal, or 20%, chance of occurring. Assuming that random number generator of the list developmental computer system can sequentially provide more or less randomly selected integer numbers in a numerical generator range between 1 and 100, i.e. a random number, r, in a sequence thereof issued from the generator is limited in the generator range to being an integer from 1 < r < 100. Then numerical passenger integer subranges for each possible passenger event are assigned as follows: 1 < r < 20: there is one additional passenger on car 1,

21 < r < 40: there are two additional passengers on car 1, 41 < r < 60: there are three additional passengers on car 1, 61 < r < 80: there are four additional passengers on car 1, 81 < r < 100: there are five additional people passengers on car 1. Operating the random number generator of the list developmental computer system to issue a more or less randomly selected integer number results in the issued number being in one of these passenger event subranges (together matching the generator range) and so selecting the number of additional passengers taken to be aboard car 1 as it leaves the lobby. Thus, if the random number generator of the list developmental computer system issues the number 23, there will be two additional passengers aboard car 1 as it leaves the lobby for a total of five passengers being aboard.

The destinations of the additional passenger assigned as being aboard CAR_k after the floor stop are then determined by the list developmental computer system on the basis that any previous car call selected destinations remaining pending after that CARk floor stop, or any new car call selected destinations immediately following such a floor stop, must be, at most, the possible floor destinations sought by any additional passengers determined to be present in that car upon leaving that floor stop in performance block 25 both in number of destinations and the locations thereof. This is so since any such additional passengers are just those without a corresponding car call because of the method followed in determining those numbers in performance blocks 24 and 25. However, here, too, this result for the maximum number, and distribution, of possible floor destinations sought by any additional passengers can then be checked for consistency with the recorded events for the next floor stop for CARk and adjusted accordingly if need be. This determination is undertaken in a subsequent performance block, 26, in Figure 1C. The transition path from Figure IB to Figure 1C is indicated by a transition balloon, B, in each figure. Following this floor destination maximum number and floor distribution determination, each of the possible added passenger destinations from one to the maximum number possible found by the list developmental computer system for CAR_k is taken to be a destination event. A corresponding numerical weight (probability) for that destination event is judged and assigned thereto based on an assumed weight (probability) distribution such that the sum of the weights (probabilities) for all such events equals 1. Another, or the same, random number generator in the list developmental computer system that can again provide a sequence of numbers more or less randomly, but limited to some numerical generator range length, is then selected to be used. Each possible destination event for CARk at the floor stop is assigned to a corresponding one of a sequence of subranges of numbers in the random number generator range with each being of a relative numerical length corresponding to the magnitude of the associated event probability, as further indicated in block 26, so that the sequence of these subranges together matches the random number generator range length. This random number generator is then operated by the list developmental computer system to more or less randomly issue a number which will thus be in one of the destination event subranges to thereby select that destination event and so select the destinations of additional passengers to be added to the further developed traffic list as indicated in a following performance block, 27. In the example scenario, following the calculation of the number of additional passengers boarding car 1 at the stop floor, the likely destinations are then determined for each of those additional passengers. Again using the lobby departure example for car 1 as was used above in determining the number of possible additional passengers, the car stop list of Table 7 shows car 1 initially leaving the lobby (or floor 0) having three car call selected destinations pending as it leaves. On this information alone, and again assuming the destination event probabilities to be equal, i.e. a discrete uniform probability distribution, (although, again, any discrete probability distribution can be used, as desired) each of the three passenger event possibilities for each of the three additional passengers has an equal, or 33%, chance of occurring. That is, each additional person has a 33% chance of going to floor 5, a 33% chance of going to floor 7 and a 33% chance of going to floor 8.

However, also in determining the added passenger destinations, the car load weight information must again be considered. Thus, here, the car stop list of Table 7 shows having additional passengers in car 1 going to floor 8 is incorrect to because the load weight shows that only zero or one passenger was on board car 1 when that car leaves floor 7 which would be the passenger boarding car 1 in the lobby that there selected that floor. Therefore, again on assumption that the destination event probabilities are equal, each additional passenger in car 1 after the floor stop in the lobby has a 50% probability of going to either floor 5 or floor 7.

Thus, again assuming that random number generator of the list developmental computer system can sequentially provide more or less randomly selected integer numbers in a numerical generator range between 1 and 100, then again that random number r in a sequence thereof issued from the generator is limited in the generator range to being an integer from 1 < r < 100. Hence, numerical passenger integer subranges for each possible destination event are assigned as follows: 1 < r < 50: an additional passenger on car 1 intends to go to floor destination 5,

51 < r < 100: an additional passenger on car 1 intends to go to floor destination 7. Operating the random number generator of the list developmental computer system to issue a more or less randomly selected integer number results in the issued number being in one of these destination event subranges (together matching the generator range) and so selecting the floor destination of an additional passenger taken to be aboard car 1 as it leaves the lobby. The random number generator is then operated again for each of then remaining randomly added passenger to assign a floor destination thereto. Continuing the lobby car departure example from above, the first random number generator random number issued was taken to be 23 so that there were two additional passengers, for a total of five passengers, on car 1 as it leaves the lobby. If, in continuing this example, the random number generator next issues additional random numbers of 47 and 3 results in both additional passengers being assigned a floor destination of floor 5 because both 3 and 47 are below 50.

The example scenario uses, in the foregoing, discrete uniform probability distributions. However, in certain situations, one of the other probability distributions available should be chosen that has its event probabilities provided more consistently with the expected outcomes in those situations. For example, if it is known that a particular floor of a building is a high user traffic floor, such as a central conference room floor, fitness club floor or cafeteria floor, then a probability distribution for passenger destination events that would make the chances of a destination being set to that floor greater than the chances of the destination being set to one of the other floors should be chosen for use with performance block 26. Similarly, if the passenger traffic in the building is known to usually be relatively large, then a distribution that makes the likelihood of more additional passengers being selected greater should be chosen for use in connection with performance block 24 rather than one giving the likelihood of fewer additional passengers.

The completion of the determination of the number of additional passengers to be added in CAR_k at each of its floor car stops by the foregoing random determination method, and the completion of the determination of floor destinations for each such added passenger also by the foregoing random determination method, allows expanding the single passenger per hall call supposition passenger list has been developed for CAR_k with these additions to thereby form a CAR_k passenger list as indicated in the next performance block, 28. The entry for each added passenger has a passenger arrival time supplied by the assignment time of the hall call responded by CAR_k for that floor stop (or, if a distribution of passenger arrival times after initial hall call was made is developed, then the selected corresponding post hall call distribution time of that hall call for that passenger is used instead). The origin for each added passenger is the floor location from which that hall call was made, and the destination supplied by the corresponding randomly determined destination.

Completing such a CAR_k passenger list for that car leads the list developmental computer system to determine if that CAR_k was the last of the elevator system operating cars for which such a passenger list is to be developed by determining in a decision diamond, 29, whether k equals N that designates the last of those operating cars in the system. If not, the list developmental computer system increments the value of k by 1 to k + 1 in another performance block, 30, and then repeats the process leading to car passenger lists by beginning another such list for CAR_k+i through returning the list development process to ahead of performance block 16 in Figure IA where it becomes the new CAR_k albeit with k now one integer greater in value. In doing so, the transition path from Figure 1C to Figure IB is indicated by a transition balloon, C, in each of those figures, and the transition path from Figure IB to Figure IA is indicated by another transition balloon, D, in each of those last two figures.

Table 9 shows one possible outcome out of the many possible for passenger lists for car 1 and car 2 in the example scenario with "NA" representing newly added passengers at the floor stop corresponding to the preceding hall call. Although, as seen in Table 2, the recorded times from the single passenger per hall call supposition passenger list of Table 8 for cars 1 and 2 corresponding to the other entries in a record in Table 9 are not parts of typical traffic, or passenger, lists, these recorded times have been retained in Table 9 to aid in understanding the method of their constructions. Here, too, since the foregoing data manipulations will be repeated for each operating elevator car in the elevator system of interest, the results are shown in Table 9 for both of the cars in this example scenario.

Table 9: One Possible Per Car Passenger List for Each Car in the Example

Scenario

Car l

Once a CAR_k passenger list is completed for each of the operating cars, each CAR_k with 1 < k < N, a building elevator system passenger list is formed as indicated in a final performance block, 31, in Figure 1C. Each passenger, rather than being identified with a recorded time or an "NA" designation, is in this list identified instead with a passenger count number, and the corresponding passenger arrival times from the CAR_k passenger lists are used to chronologically order the passengers to then which a sequential count number is assigned. The record corresponding to each passenger, or count number, in the list contains a field for the corresponding arrival time for that passenger, a field for the floor of origin of that passenger, and a field for the floor of destination of that passenger just as do the individual CAR_k passenger lists.

Table 10 shows one possible building elevator system passenger list out of many possible for the example scenario based on the probabilistically determined CAR_k passenger lists of Table 9 in the possible outcome for such lists shown there. There is repeated in Table 11 the exact, or actual, traffic, or passenger, list from Table 2. While the lists of passenger records in the two tables 10 and 11 are similar, these lists are nonetheless different. This difference is to be expected due to the probability based determinations made in developing the list in Table 10. However, there can typically be expected to be a greater similarity between Tables 10 and 11 than between Table 10 and the single passenger per hall call supposition passenger list of Table 8 which lacks any representation of the presence of additional passengers boarding at the same floor as others who first entered car calls after boarding that these additional passengers accepted as their destinations also.

Table 10: One Possible Building Elevator System Passenger List in the Example Scenario

Table 11: Exact Passenger List for Example Scenario (Repeat of Table 2)

Although the present invention has been described with reference to preferred embodiments, workers skilled in the art will recognize that changes may be made in form and detail without departing from the spirit and scope of the invention.

Claims

CLAIMS:

1. A method for determining a traffic list for elevator cars operating in an elevator system in a building having a tabulation of passengers chronologically listed with arrival times thereof at corresponding floors of origin for transport to floors of destination selected by them in those operating elevator cars and in which system selected performance events associated with those elevator cars and the corresponding times of occurrence thereof can be recorded in a data storage facility during operations of those elevator cars over selected time durations, the method comprising: acquiring from the data storage facility a chronologically ordered list of operating elevator car performance events and the corresponding times of occurrence thereof; and forming an initial passenger list based on assuming there is a single passenger in each of the operating elevator cars departing from origin building floors after arriving there in response to each summons made in the system from those floors for transportation service to other building floor destinations, and adding thereto a number of passengers in each of the operating elevator cars following such departures thereof equal to a number of requests made in each such elevator car, beyond one such request, for transportation service to a selected building floor destination prior to those elevator cars subsequently arriving at a next building floor.

2. The method of claim 1 further comprising: determining from the chronologically ordered list of operating elevator car performance events and the corresponding times of occurrence thereof whether each of the operating elevator cars in each subsequent departure from a building floor, after arriving at that building floor in response to a summons made in the system therefrom so as to have that building floor as the origin building floor, could possibly have had, upon any of those subsequent departures, additional passengers therein beyond those represented as being therein in the initial passenger list and what number of additional passengers are possible in each operating elevator car having such a possibility of one or more additional passengers therein at those subsequent departures thereof to thereby be a possibility departure; and determining through using a chance based procedure, for each possibility departure of each operating elevator car, the number of additional passengers, out of the number of additional passengers possible, to be assigned to that operating elevator car at that possibility departure.

3. The method of claim 1 wherein the acquiring from the data storage facility of a chronologically ordered list of elevator car performance events and the corresponding times of occurrence thereof results in a tabulation including hall call events made in the system from building floors by hall occupants to request elevator car transportation service to another building floor destination along with the origin building floor of each hall call, building floors arrival and departure events of the operating elevator cars, car call events made in the system from within operating elevator cars to request transportation service by that elevator car to a selected building floor destination along with the destination floor selected in each car call, and load weight indications for the operating elevator cars immediately following each departure event thereof from a building floor.

4. The method of claim 1 wherein the forming of an initial passenger list has passenger records provided therein through pairing each request made in each such elevator car with the most recent summons made in the system from a summons origin building floor assigned to that elevator car for response thereby through arriving at that building floor, and with each passenger record having an arrival time therein corresponding to a time of that most recent summons and an origin floor that is the summons origin floor, and further a destination floor that is the selected building floor destination for that paired request.

5. The method of claim 2 further comprising: determining the number of pending requests for transportation service to a selected building floor destination at each possibility departure of each operating elevator car that were made prior to those elevator cars subsequently arriving at a next building floor; determining through using a chance based procedure, for each additional passenger assigned to an operating elevator car at a possibility departure thereof, which of the building floor destinations in the pending requests is to be assigned to that additional passenger; and forming a chance based passenger list by adding to the initial passenger list those additional passengers assigned to each operating elevator car at a possibility departure with each additional passenger having a passenger record therein including an arrival time corresponding to the time of the summons from a summons origin floor to which that operating elevator car responded by arriving at that summons origin floor immediately prior to this possibility departure and an origin floor which is the summons origin floor, and with a destination floor that is that building floor destination assigned to that additional passenger.

6. The method of claim 2 wherein the determining of the situation of each of the operating elevator cars in each subsequent departure from a building floor, after arriving at that building floor in response to a summons made in the system therefrom so as to have that building floor as the origin building floor, as to whether that elevator car could possibly have had, upon any of those subsequent departures, additional passengers therein beyond those represented as being therein in the initial passenger list and the number of additional passengers possible is done so as to be consistent with situations for that elevator car also at adjacent in time arrivals and departures at building floors including the weight load indication in these situations.

7. The method of claim 2 wherein the determining through using a chance based procedure of the number of additional passengers to be assigned to an operating elevator car at a possibility departure comprises assigning a numerical weight to each of the numbers of additional passengers possible so that these weights together total one and using a random number generator to select the number of additional passengers to be assigned to that elevator car at that possibility departure.

8. The method of claim 5 wherein the determining of the situation of the number of pending requests for transportation service to a selected building floor destination at each possibility departure of each operating elevator car is done so as to be consistent with situations for that elevator car also at adjacent in time arrivals and departures at building floors including the weight load indication in these situations.

9. The method of claim 5 wherein the determining through using a chance based procedure which of the building floor destinations in the pending requests is to be assigned to each additional passenger assigned to an operating elevator car at a possibility departure thereof comprises assigning a numerical weight to each of the numbers of building floor destinations possible so that these weights together total one and using a random number generator to select the building floor destination to be assigned to each additional passenger.