EP1735229A2

EP1735229A2 - Method for controlling an elevator system

Info

Publication number: EP1735229A2
Application number: EP05730844A
Authority: EP
Inventors: Marja-Liisa Siikonen; Jari Ylinen
Original assignee: Kone Corp
Current assignee: Kone Corp
Priority date: 2004-04-15
Filing date: 2005-04-12
Publication date: 2006-12-27
Anticipated expiration: 2025-04-12
Also published as: ES2340689T3; EP1735229B1; WO2005100223A2; DE602005019866D1; WO2005100223A3; FI20040544A0; FI115396B

Abstract

The present invention discloses a method and a system for the allocation of elevators on the basis of calls in an elevator system. The method produces for different route alternatives a cost function wherein the traveling time of each passenger is calculated. The calculation of the traveling time is performed taking into account the time spent while waiting for an elevator at a floor, the ride time in the elevator car, the delaying effect of active landing and car calls on the traveling time and the delaying effect of car calls given by new passengers entering at intermediate stops. The method is advantageous in the destination call system, but in the traditional up-down call system it is possible to utilize information provided by traffic statistics to predict the destination floor. Route alternatives can be created e.g. by using genetic algorithms. Once the route alternative giving the shortest average traveling time has been calculated, the elevators are controlled according to this route. In intensive traffic conditions, traveling time is optimized, and in quiet traffic conditions, passenger waiting time is optimized.

Description

METHOD FOR CONTROLLING AN ELEVATOR SYSTEM

FIELD OF THE INVENTION

The present invention relates to control of an eleva- tor group .

BACKGROUND OF THE INVENTION

An elevator system can be controlled by two different principal methods, of which the more traditional and more widely used method is up-down call buttons at the elevator landing floors and a car call panel inside the elevator car. This traditional call system requires that the elevator passenger give two successive calls: a landing call (ordering an elevator to the particular departure floor) and a car call (indicating the target floor to the elevator system) . In this call system, the elevator to serve the call can be announced either immediately after the elevator control system has allocated the call (decided which elevator is to serve the call), or e.g. only after an arriving elevator starts braking to stop at the departure floor of the person having issued the call .

The other call system is so-called destination con- trol, in which the elevator customer gives only one call. The call is given like a car call at the elevator landing floor by inputting destination floor information via a floor button panel or e.g. using a numeric keypad. In the destination call system, the al- location of elevators can be accomplished in a more sensible way because the system learns the information relating to each passenger (departure floor and destination floor) at an earlier stage and the passengers' destination floors can already be taken into account when a suitable elevator is being allocated. In the case of large elevator systems and large numbers of passengers, it is thus possible e.g. to assign the same elevator to customers traveling to the same floor.

One of the functions of the elevator system is to allocate elevators to calls in such a. way that a desired cost function can be minimized. The cost function may comprise summed passenger waiting times, traveling times, electric energy consumption of the system, num- bers of times the elevator car has stopped at different floors, or the aforesaid or other desired quantities may be weighted with desired weighting coefficients .

An effectively working elevator control algorithm required in large buildings is a very complicated optimization process. For example, the traditional collective control works in such manner that a given landing call is allocated to an elevator ttiat is traveling to- wards the floor from which the landing call was input and is located closest to the call input floor. On the other hand, this leads to an accumulation of elevators with several elevators moving in th.e same direction as a front. As a further consequence of this, the overall performance of the elevator system deteriorates.

In prior art, the most optimal ele-vator is found e.g. by the ESP method (Enhanced Spacing Principle) . In ESP, the issued calls are observed and the passenger waiting times are optimized. The number of passengers associated with each landing call and waiting on the floor in question is estimated as far as possible on the basis of statistical data. Those landing calls that the system assumes to be associated with the largest number of elevator customer's are served fastest . Another method for allocating elevators on the basis of calls is to use genetic algorithms, especially in large elevator systems. Genetic algorithms are described e.g. in patent specification FI112856B. Ge- netic algorithms do not guarantee that the absolutely most optimal value is found, but the results obtained in practical applications are quite close to it. In genetic algorithms the routes of the elevators in the system can be encoded in different chromosomes, in which one gene defines an elevator customer and the elevator to serve him/her. The system starts the process e.g. from a randomly selected route alternative and applies to it various genetic procedures, such as proliferation, crossbreeding and mutation. One genera- tion at a time, a number of new chromosomes are generated via these genetic procedures and at the same time the chromosomes thus obtained are examined to determine whether they are viable for further processing. Viability may mean e.g. that the waiting time falls below a given value. Crossbreeding means that two route alternatives are combined at random to create one new route alternative. In mutation, the values of the genes of the chromosome are changed arbitrarily. The chromosome results given by the algorithm converge at some stage, and from the last processed set of chromosomes the one having the best viability is selected. The passengers are allocated to elevators according to the genes of the best chromosome.

In connection with the present invention, the starting point is optimization of total journey time. In prior art, an allocation principle is described in patent specification FI82917C. According to this method, the sum of so-called internal an external service costs is calculated. Internal service costs here refer to the waste time spent by passengers in the elevator car due to intermediate stops, and external service costs re- fer to the wait ing time spent by passengers in an ele vator lobby . Patent specification FI82917C mentions a cost funct ion K :

K = t_v(P_m + k_l ^■ R_E - k₂ - R_c) + k₁[m - t_m +t_v(R_E + R_c -R_EC + Z)^ , (1 )

where t_v is delay time during an intermediate stop, P_m is instantaneous load at the instant of calculation, R_E is the number of intermediate floor calls issued between the floor of current location of the elevator and the floor where the customer is to be picked up, R_c is the numbeir of car calls given between the floor of current location of the elevator and the floor where the customer is to be picked up, ki is the num- ber of passengers entering the elevator for one landing call as estimated on the basis of the prevailing traffic situatzLon, k₂ is the number of passengers leaving the elevator for one car call as estimated on the basis of tlie prevailing traffic situation, m is the number of floor-to-floor intervals between the floor of current location of the elevator and the floor where the customer is to be picked up, t_m is average journey t±me for one floor-to-floor interval, R_EC is the number of coincident car calls and lainding calls between the floor of current location of the elevator and trie floor where the customer is to be picked up, Z is a additional factor depending on the operational condition of the elevator car, and where the first term of the sum represents internal service costs and the second term represents external service costs. Thus, the cost function optimizes the waiting time, which is obtained as the sum of the waiting time spent in the lobby and the waste time spent in the elevator car due to stops. Coincident calls (which here means that an active landing call addressed to the elevator is simultaneously a destination floor given as an active car call) are taken into account in the method .

Patent specification US4991694 deals with immediate allocation of destination calls. This specification defines the cost function K as follows:

K = K„ κ„ +κ p„s +κ„ +κ.. +κ„ (2)

where K_rs is tb_e waiting time of new passengers at the call input floor, K_rz is the traveling time of new passengers, K_ps is the length of the waste time spent by car passengers due to an intermediate stop caused by a landing call, -_pz is the length of the waste time spent by car passengers due to an intermediate stop caused by a car call, K_ws is the waiting time of all passengers entering the elevator between the call input floor and the destination floor, and K_W2 is the waiting" time of all passengers entering the elevator after ar- rival at a floor requested by an active destination floor call. In this method, the costs are optimized on the basis of passenger waiting times. As stated above, waiting time is accumulated from waiting in the elevator lobby and from intermediate stops due to landing calls and car calls.

The indexing msed in patent specifications FI82917C and US4991694 is a call-based arrangement, i.e. all calculation in the algorithm is performed on the basis of calls. This produces inaccurate results because it is possible th-at two or more passengers arrive in an elevator lobby at the same time but only one of thera gives a landing call (up or down call) to the elevator system. In sucti a case the system assumes that only one passenger is arriving to the elevator and selects a suitable ele^~vator accordingly. If the system knew that, for example, three passengers arrive to the ele- vator on a given floor, then, depending on whether the passengers are going to th-e same destination floor or whether they are all going to different floors, this would have an effect on ttie cost function and at the same time the allocation result would probably be different in these two example cases .

Thus, the problem with prior-art solutions is that the algorithms used are based on calls issued, which is a source of inaccuracy in the elevator allocation result to be calculated. In addition, prior-art elevator control algorithms optimize passenger waiting times in the elevator lobby and car and do not take the actual traveling time in the elevator car accurately into ac- count. Precisely speaking, prior-art methods do not optimize passengers' traveling time, because they do not take the time losses resulting from the car calls given by new passengers entering the car at intermediate stops accurately into account.

OBJECT OF THE INVENTION

The object of the present invention is to overcome some of the above- mentioned problems in elevator control. The aim is to create a control method in which both the waiting time and the traveling time of passen- gers are optimized.

BRIEF DESCRIPTION OF THE INVENTION

As for the features of the present invention, reference is made to the claims.

The present invention deals with a method for allocating elevators on the basis of call data, and the method is especially intended for use in a destination call system, wherein both the source floor and the destination floor of the customer are already known after the customer has given a call in an elevator lobby. Source floor refers to the floor where the customer gives a landing call or destination call and where the customer en- ters the elevator. The source floor is thus the same as the customer's departure floor. Based on active calls and the location and operational condition of the elevators at the instant under consideration, all possible elevator route alternatives are calculated.

In the method, a cost function is calculated wherein passenger-specific average total traveling time, i.e. the time from the instant the person gives a destination floor call to the instant he/she leaves the elevator at the destination call floor, is minimized. The procedure takes into account the waiting time spent at the elevator landing floor, besides the traveling time spent in the elevator car as well as the delays caused by intermediate stops that, as far as known, are to be made during the journey. Intermediate stops may be due to active destination floor calls or source floor calls given by new passengers along the route of the elevator. Further delays arise in consequence of destination floor calls given by new customers boarding at intermediate stops. The calculation is performed taking into account the time it takes the elevator passenger to move from the OOP panel (Destination Operating Panel = device for inputting a destination floor call) to the elevator lobby, from the lobby to the elevator or vice versa, the time elapsing while the elevator door is being opened and closed, the waste time caused by the braking prior to stopping of the car and the acceleration after the stop, and additionally possible other waste time due to mechanical or electrical delays within the elevator system or guard times related to operational safety.

In intensive traffic conditions, elevators are reserved for passengers according to the route alternative that gives the shortest average traveling time, whereas in quiet traffic conditions the average passenger waiting time is optimized. Once the calculation result is obtained, the elevator customer is immediately informed as to the elevator allocated to serve him/her.

The method of the present invention can be combined with the use genetic algorithms to determine the most advantageous route alternative. However, the route alternatives processed by the algorithm can also be created in other ways. If only traditional landing call buttons and a car call panel are in use, it is necessary to weight the traveling time associated with a given landing call by the relevant predicted number of passengers . Further, measured traffic statistics can be utilized to estimate passengers' destination floors at a given instant of time from a given source floor. Ttie results of the forecast can be further utilized when the algorithm of the present invention is used.

A feature characteristic of the method of the present invention is that it employs passenger-specific calculation instead of call-specific calculation.

As traveling time is minimized and, in a preferred embodiment, the destination call system is used, the capacity of the elevator system can be better utilized as compared to the control algorithms used in the traditional up-down call system. Another significant advantage of the present invention is that the same control system can be used to control both systems using destination calls and systems using traditional up-down calls. In the destination call system, the traveling times are optimized and the serving elevator is immediately signaled to the customer. In this operating mode and in peak traffic conditions, the number of intermediate stops can be effectively reduced and the elevator capacity can be more efficiently utilized. Immediate signaling can also be used in a system comprising up- down call buttons. The signaling can be given automatically or it can be set manually to a suitable value with respect to usability.

The control system also permits the destination operating mode to be set into an active state e.g. only during peak traffic hours while at other times traditional call buttons are in operation. As another alternative, the desti- nation operating mode may be continuously in use.

LIST OF FIGURES

Fig. 1 presents the components of a cost function generated in elevator control according to the present invention, and

Fig. 2 presents the components of an elevator system associated with the present invention. DETAILED DESCRIPTION OF THE INVENTION

In a preferred embodiment, the present invention is applied in -so-called destination call control, wherein the elevator system control receives the information regarding the customer's source and target floors at an early stage. However, the present invention is also applicable for use i n a traditional elevator system provided with up-down call buttons.

The cost function is first presented in the case of destination call control and then in the case of traditional call control. Fig. 1 presents in a simple form the time terms needed in the generation of the cost function .

In the present invention, the object of optimization 11 is the passenger's average traveling time. The traveling time contains the passenger's wait- ing time on his/her source floor 10, which means the time interval from the input of a landing call to the arrival of the elevator car. Further, the traveling time contains the ride time in the elevator car 12. In addition to these, it is necessary to take into account the landing and car calls 13 that the system knows of at the instant of input of the customer's destina- tion call and that are to be allocated to be served by each elevator under consideration. These are a source of delay in the time spent by the elevator customer in the elevator car. Moreover, the car calls 14 given by new passengers boarding the elevator car have to be taken into account in the traveling time (provided that the system already knows of these calls as destination calls), because they have a similar effect of increasing the traveling time of the passengers already in the car.

In the optimization method of the present invention, a cost function 11 is created which, based on the above considerations, can be expressed in a simple form e.g. as

J„ = m(JT_{f enlerJift} + JT_incM ) , 3)

where J_av is the passenger's average traveling time when a new passen- ger uses the optimal elevator. JT_f__enter,ii_ft is the sum of passenger waiting times associated with a given destination floor or, in the case of a desti- nation call, the traveling time of the person having input the call . JTj_nCιij_ft is the length of the waste time that, for different reasons, is summed in the traveling time of the passengers riding in the car.

If the call data is available immediately in connection with the input of the landing call (in the case of a destination call), optimization 11 can be performed immediately and the most advantageous elevator obtained as a final result 15 of the optimization can be notified to the elevator customer. If the elevator is originally located at floor 'liftpos' (which is different from the elevator customer's source floor), the journey time for a passenger starting a ride on the elevator in question from floor 'f_enter' and giving floor 'f exit' as the destination floor will be obtained as follows:

f_enter,lift ~ ^^ -"-liftpos, f _enter,lifl ^ ^■^^■'^■^-f _enler,f_exit,lifl = jOUmey time, ' '

where ETAiift_Pos,f_enter,iift is the waiting time of the new passenger or passengers at the entry floor and ETAf__enter,f_eχiUift '^s the ride time- from the entry floor to the destination floor.

ETA_xι_ιX2,ii_ft is the estimated journey time for the elevator (Estimated Arrival

Time) from floor x to floor x₂: .outjifi ^{' ■}' pass ^{+ •} door, lift ^"*^"" * acc.dec ) >

(5) where T_xi_>X2,_driv_e is the elevator journey time at a constant travel speed from floor _x to floor x₂. The sum term represents the extra time resulting from stops due to landing and car calls, which is spent during the trip before floor x₂ is reached. Fj j_ft = 1 if the elevator stops at floor i, otherwise Fj jft = 0. Njjn ift and Nj_ι0U.,iift are the numbers of passengers entering and leaving the elevator, respectively. T_paSs is the average time required for a passenger to step into or out of the elevator. T_door.nft is the additional time consumed by the door opening and closing operations, and T_acc,dec represents the delay resulting from the acceleration and braking of the elevator as compared to travel at an even travel speed. The traveling time of the elevator customers already riding in the elevator car is increased by new elevator customers giving new landing calls (destination calls) on their entry floor and by the stops required to drop these new customers off at their destination floors when the destination floor is between x-i ... x₂. The magnitude of this additional delay caused by new passengers to those already riding on the elevator is

J incjift ~ -'f_ enter, lifl V pass ^"*^" f _ enter, lift door, lift + ^ acc,dec // ^{" "} _lr..

. (⁶)

^•^ ' _exιt,lιft pass ^"*^" f _exιtjιfi V c door, lift + τ ace, dec ))

where Lf__enterjift and Lf__eXjtjift are elevator's loads (=numbers of passengers) at the floors where a new passenger either boards or leaves the elevator. Yi.ii_ft = 0 if the elevator would have stopped at floor i without the new passenger giving a call, otherwise Y^m = 1.

In formula (6), it is necessary to consider possible coincidences, i.e. cases where a destination call given by a first passenger and a landing call given by a new passenger have the same floor as a destination and are allocated to the same elevator. In addition, in formula (6) the calculation of the traveling time is weighted by the number of passengers on the elevator in such manner that e.g. in a full elevator a significant extra delay is produced if a new landing call is allocated to be served by such a full elevator.

Let us now consider a traditional elevator system provided with up-down call buttons and a car call panel. In an elevator system provided with normal up-down call buttons, the traveling time can be calculated for example by assuming that each landing call corresponds to one elevator customer. The waiting times obtained by considering each call are summed. However, this produces quite unsatisfactory results when an allocation algorithm is used, because it is quite usual that the same land- ing call represents several elevator customers queuing up to enter the elevator. In a traditional elevator system like this, if the number of passengers associated with one call can be determined in some way, then this can be utilized in the calculation of the sum of waiting times.

In the estimation of traveling times, it is possible to use traffic statistics for the building in question. From the statistics it is possible to find out which floor the passenger most typically rides to at a given time on a given day of the week when additionally the source floor is known. Similarly, to assist the calculation, e.g. the average traveling time can be determined from the statistics when the time of the day and the source floor are known. In this way, the calculation can be executed efficiently immediately upon input of a landing call in a traditional system provided with up- down call buttons and it is not strictly necessary to await a destination call given by the customer in the elevator car.

The summed total journey time for passengers arriving to the elevator upon a single landing call is:

f _enter, lift, total ⁼ 2-_l

= total journey time

where P is the number of passengers departing from floor f_enter upon a single landing call.

The additional term JT|_nc i_ft contains the additional time consumption caused by new passengers in the traveling time of the passengers al- ready in the car:

J incjift ~ ^■L'f _ enter, lift f_ enter, lift-' pass ^"■" "* / _enterJιft V door, lift ^"*^" ' acc.dec )~*~ f^' / Yi (8)

2-ι L Uift lifl pass ^~ ι,lιfi door, lift ⁺ acc.dec ]\ i-f _enter

where P_f enterjift is the number of new passengers entering the elevator from a given floor and Pjjj_ft is the number of passengers leaving the elevator from the number that boarded the elevator at floor f_enter. Kj i_ft = 1 if at least one passenger having entered from an 'intermediate floor' leaves the elevator at floor i and the car had not yet stopped at this floor before. Otherwise Kiji_ft = 0. Thus, in the sum term practically all destination call floors given by new passengers are considered and the delays occurring when these passengers leave the elevator are summed taking care that each door operation and braking/acceleration of the car is taken into account only once. The cost function (3) has now been defined for both the destination call system and the traditional call system. During intensive traffic, the cost function minimizes the average passenger-specific traveling time, which 5 comprises the time spent while waiting for an elevator, the actual ride time and additionally the delays caused by passengers subsequently entering the elevator. Once the minimum has been found, each passenger is directed to his/her right elevator according to the elevator allocation consistent with the shortest traveling time. The elevator system control0 naturally performs calculations continuously so that new calls entered and the continually changing positions of the elevators in the system are properly taken into account in the control of the elevators. Since the total traveling time is the object of optimization in the case of intensive traffic, the elevator capacity can be effectively reused after the customer's eleva-5 tor trip.

During quiet traffic, the algorithm minimizes passenger-specific average waiting time. In quite traffic conditions, the elevator arrives quickly to the call input floor, but the elevator is allowed to make even several interme- o diate stops if necessary.

The car loads are balanced by the algorithm so that the given car load limits are not exceeded. The control method allows the cars to be filled to the upper limit of the number of persons if people enter the elevator from5 the same source floor. In practice, this limit is only reached when a special peak traffic condition prevails in the system. Peak traffic again can be identified e.g. from measured statistical traffic data or from traffic forecasts made.

o Fig. 2 presents an example of an actual elevator system employing the above-described method, showing the essential parts of the system. The building is provided with an elevator system comprising elevators 20. The call input equipment 21 includes both traditional up-down call buttons and a car call panel placed in the car. Furthermore, the call input equipment 5 21 contains the buttons required in a destination call system on each floor. The intelligence of the system is located in a control system 22 comprising a microprocessor (not shown in the figures) as an essential part of it. The microprocessor contains a memory, in which a computer program capable of executing the method of the present invention (or a part of it) is stored. The memory may also be implemented as an external part connected to the computer. The microprocessor runs the program code comprised in the computer program, thus executing the various stages of the method of the present invention (or part of them).

In the form of the program code executed in the microprocessor, the traveling time is calculated by a time counter 23. The control system 22 performs the required optimization operations by using the input data and method of the present invention. Previously measured traffic statistics 24 can be utilized when an optimization algorithm is used. The traffic statistics 24 may be stored in a separate memory block. As a final result, the control system 22 calculates the optimal elevator route alternative that minimizes the average traveling time.

The invention is not limited to the embodiment examples described above; instead, many variations are possible within the scope of the inventive concept defined in the claims.

Claims

1. A method for the allocation of elevators to passengers in an elevator system, in which method the source and destination floors of each passenger is made known to the elevator system via a call input

5 equipment, c h a r a c t e r i z e d in that the method comprises the steps of: forming the possible elevator route alternatives on the basis of active calls and the state of the elevators at the instant under consideration; determining for the elevator route alternatives a waiting time from o the input of the source floor call to the arrival of the elevator at the call input floor and a ride time in the elevator car from the source floor to the destination floor; determining for the elevator route alternatives a first delay caused by intermediate stops to be made between the source and destination5 floors of the passengers; determining for the elevator route alternatives a second delay caused by intermediate stops to be made due to destination floor calls given by customers having boarded the elevator between the source and destination floors of the passengers; o determining for the elevator route alternatives a passenger traveling time based on the waiting time, the ride time, the first and the second delays; and allocating the elevators to the passengers according to traffic intensity in accordance with the route alternative that gives the shortest waiting5 or traveling time.

2. A method according to claim 1 , c h a ra c t e r i z e d in that the aforesaid first delay results from landing calls allocated to the elevator and car calls given in the elevator car.

3. A method according to any one of the preceding claims 1-0 2, c h a r a c t e r i z e d in that the method further comprises the step of: determining the aforesaid first and second delays by summing the average time it takes for each elevator passenger to move from the floor to the elevator or vice versa, the waste time caused by opening and closing of the elevator door, the waste time caused by braking and accelera-5 tion of the elevator and the waste time caused by the operation of the system.

4. A method according to any one of the preceding claims 1- 3, characterized in that genetic algorithms are used in the allocation of the elevators.

5. A method according to any one of the preceding claims 1- 4, characterized in that the aforesaid source and destination floor data are given to the system via a destination call system (DOP).

6. A method according to any one of the preceding claims 1- 5, characterized in that the aforesaid source floor data is given to the system via landing call buttons and the destination floor data via a car0 call panel.

7. A method according to claim 6, characterized in that the method further comprises the steps of: calculating the traveling time associated with one landing call by multiplying the defined traveling time by a predicted number of passen-5 gers associated with one landing call; and predicting the passengers' destination floors on the basis of traffic statistics.

8. A method according to any one of the preceding claims 1- 7, characterized in that in intensive traffic conditions the passen- 0 ger's average traveling time is optimized and in quiet traffic conditions the passengers' average waiting time is optimized.

9. A system for the allocation of elevators to passengers in an elevator system, said system comprising at least one elevator (20); 5 a call input equipment (21) for making the source and destination floors of each passenger known to the elevator system; characterized in that the system further comprises: an elevator control system (22) for forming the possible elevator route alternatives on the basis of the calls active and the state of the ele- o vators at the instant under consideration; a time counter (23) for determining for the elevator route alternatives a waiting time elapsing from the input of the source floor call to the arrival of the elevator at the call input floor and a ride time elapsing in the elevator car from the source floor to the destination floor; 5 a time counter (23) for determining for the elevator route alternatives a first delay caused by intermediate stops to be made between the source and destination floors of the passengers; a time counter (23) for determining for the elevator route alternatives a second delay caused by intermediate stops to be made due to destination floor calls given by customers having boarded the elevator between the source and destination floors of the passengers; a time counter (23) for determining for the elevator route alternatives a passenger traveling time obtained on the basis of the waiting time, the ride time, the first and the second delays; and an elevator control system (22) for allocating the elevators to the passengers according to traffic intensity in accordance with the route al- ternative that gives the shortest waiting or traveling time.

10. A system according to claim 9, characterized in that in the time counter (23) the aforesaid first delay is caused by the landing calls allocated to the elevator and the car calls given in the elevator car.

11. A system according to any one of the preceding claims 9-

10, characterized in that the system further comprises: a time counter (23) for summing theaverage time it takes for each elevator passenger to move from the floor to the elevator or vice versa, the waste time caused by opening and closing of the elevator door, the waste time caused by braking and acceleration of the elevator and the waste time caused by the operation of the system, determining the first and second delays.

12. A system according to any one of the preceding claims 9-

11, characterized in that in the elevator control system (22) ge- netic algorithms are used for the allocation of the elevators.

13. A system according to any one of the preceding claims 9-

12, characterized in that the call input equipment (21) is a destination call equipment (DOP) for the input of source and destination floor data to the system.

14. A system according to any one of the preceding claims 9-

13, characterized in that the call input equipment (21 ) consists of landing call buttons and car call buttons for the input of source and destination floor data to the system.

15. A system according to claim 14, characterized in that the system further comprises: a time counter (23) for calculating the traveling time associated with one landing call by multiplying the defined traveling time by a predicted number of passengers associated with one landing call; and traffic statistics (24) for predicting the passengers' destination floors.

16. A system according to any one of the preceding claims 9-

15, c h a r a c t e r i z e d in that in intensive traffic conditions the passenger's average traveling time is optimized and in quiet traffic conditions the passengers' average waiting time is optimized.

17. A computer program for the allocation of elevators to passengers in an elevator system, wherein the source and destination floors for each passengers are made known to the elevator system via a call input equipment, and which computer program comprises a program code, c h a r a c t e r i z e d in that the program code has been arranged to execute the following steps when run on a data processing apparatus: forming the possible elevator route alternatives on the basis of active calls and the state of the elevators at the instant under consideration; determining for the elevator route alternatives a waiting time from the input of the source floor call to the arrival of the elevator at the call input floor and a ride time in the elevator car from the source floor to the destination floor; determining for the elevator route alternatives a first delay caused by intermediate stops to be made between the source and destination floors of the passengers; determining for the elevator route alternatives a second delay caused by intermediate stops to be made due to destination floor calls given by customers having boarded the elevator between the source and destination floors of the passengers; determining for the elevator route alternatives a passenger traveling time based on the waiting time, the ride time, the first and the second de- lays; and allocating the elevators to the passengers according to traffic intensity in accordance with the route alternative that gives the shortest waiting or traveling time.

18. A computer program according to claim 17, c h a r a c - t e r i z e d in that the aforesaid first delay is caused by the landing calls allocated to the elevator and the car calls given in the elevator.

19. A computer program according to any one of the preceding claims 17-18, characterized in that, when run on a data processing apparatus, the program code has further been arranged to execute the step of: determining the aforesaid first and second delays by summing the average time it takes for each elevator passenger to move from the floor to the elevator or vice versa, the waste time caused by opening and closing of the elevator door, the waste time caused by braking and acceleration of the elevator and the waste time caused by the operation of the system.

20. A computer program according to any one of the preceding claims 17-19, characterized in that genetic algorithms are used in the allocation of the elevators.

21. A computer program according to any one of the preced- ing claims 17-20, characterized in that the aforesaid source and destination floor data are given to the system via a destination call system (DOP).

22. A computer program according to any one of the preceding claims 17-21.characterized in that the aforesaid source floor data is given to the system via landing call buttons and the destination floor data via a car call panel.

23. A computer program according to claim 22, characterized in that, when run on a data processing apparatus, the program code has been arranged to execute the steps of: calculating the traveling time associated with one landing call by multiplying the defined traveling time by a predicted number of passengers associated with one landing call; and predicting the passengers' destination floors on the basis of traffic statistics.

24. A computer program according to any one of the preceding claims 17-23, characterized in that in intensive traffic conditions the passenger's average traveling time is optimized and in quiet traffic conditions the passengers' average waiting time is optimized.

25. A computer program according to claims 17-24, c h a r- acterized in that the computer program is stored on a medium that can be read on a data processing apparatus.