Elevator Scheduling Ingrid Cheh Xuxu Liu 050509 Elevator

Elevator Scheduling Ingrid Cheh Xuxu Liu 05/05/09

Elevator Scheduling Problem l Elevator as a control system ¡Response time and behavior depends on programmed algorithm(s) ¡Different solution depending on building type and number of elevators working together l In algorithm, assignment of job: ¡External elevator request ¡Internal floor request

Our Problem Context l Setting: ¡ 15 floor school administration building ¡ 2 elevators l Goal: ¡ Analysis of possible elevator scheduling algorithms through simulation ¡ Find most optimal in given problem ¡ Suggest future directions

Tools l Basic graphical user interface & software ¡Programmed in Pascal in Delphi programming environment l Strategies, labeled A, B, C & D

User Interface & Software

Strategy A l Elevators calculate most requested floor Reacting to modal distribution of requests l Each elevator heads to most requested floor, unless: ¡ Other elevator already heading there ¡ Other elevator contains passenger who want to go there l If no passengers in elevator, elevator will position itself on floor 8

Strategy B l If no passenger in elevator, it waits and goes to the floor where earliest request is made External requests in FIFO manner l Otherwise, elevator will head to floors requested in order of passenger entry into elevator l If elevator passes a floor where current passenger has requested to get off, elevator stops and picks up new passengers on direction of travel

Strategy C l Both elevators do the same Internal before external requests in FIFO manner l If elevator is empty, it heads towards the earliest external request l Otherwise, elevator would head towards desired destination of occupants l When elevator opens, it picks up passengers in either direction

Strategy D l Elevator A Round Robin scheduling method ¡ Initially goes to floor 15 and descends one floor at a time picking up passengers who are going down ¡ Heads straight back up to floor 15 when reaches floor 1 l Elevator B ¡ Initially goes to floor 1 and ascends one floor at a time picking up passengers who are going up ¡ Heads straight back down to floor 1 when reaches floor 15

Performance Metrics Average Wait (w) Passengers Carried (c) Passengers Using Stairs (s) Metrics Passenger’s Satisfaction (p) Power Efficiency (e) Service Profile (chart) l w: time request submission to final destination arrival l c: total passengers carried l s: total passengers who gave up elevator wait and used stairs l p: c/(c+s) l e: energy efficiency level in elevator operation l chart: running profile of customers served in different time brackets

Inputs to Simulation Available Inputs l Scenario ¡ Particular pattern of passengers waiting and choice of floor l Peak Hours ¡ Choice of more or less passenger traffic l Simulation Length ¡ Total simulated of each simulation run l Boredom Level ¡ How much time before switching to stair use Our Selection of Inputs l Scenario ¡ 30 different scenarios l Peak Hours ¡ Peak chosen l Simulation Length ¡ 5 minutes l Boredom level ¡ Level of 20

Results 73. 70 41. 07 84. 27 61. 40 75. 93 34. 77 79. 93 34. 57 24. 67 19. 27 12. 07 19. 73 38. 90 29. 70 27. 60 61. 20 57. 53 20. 33 70. 57 57. 27

Analysis of Results l Strategy A is the best strategy ¡Least average waiting time ¡Least number of passengers using stairs ¡Maximum passengers carried ¡Greatest percentage of passenger’s satisfaction l Ranking of strategies from best to worst: ¡A, C, D, B

Limitations of Strategies l Strategy A ¡Efficiency may be related to its position on 8 th floor l. Overlooks the potential heavy skew of requests around the lower floors ¡Difficult for Elevators A and B both to calculate the most requested floor l Interlacement of strategies B & C necessary l “Smart” strategies are less power efficient

Extensions within Problem Context l Investigation into other variables ¡ Simulation length ¡ Peak hours to non-peak option ¡ Boredom level ¡ Scenario l Uniform probability distribution is not realistic l Poisson arrival processes might be more reflected l More requests assigned to ground level l Different setting l ± Floors l ± Elevators

Inspiration from Literature l Passenger behavior (Susi, Sorsa and Sikonen) ¡ Modeling of diverse traffic flows with passenger compositions that incorporate physical and behavioral characteristics l Complex controllers (Bartz-Beielstein, Preuss and Markon) ¡ Fujitec ¡ Neural network structure to determine control strategy l Zoning Policy (Chu, Lin and Lam) ¡ Set of floors divided into blocks ¡ Goal of increasing overall handling capacity

Summary l Insight into elevator scheduling simulation for a particular setting l Compromise between running efficiency and power efficiency l Very simple and limited model l Extensions possible, as shown by literature research

Thank You!