S. D. Deshmukh OM V. Capacity Planning in Services u Matching Supply and Demand u The Service Process u Performance Measures u Causes of Waiting u Economics of Waiting u Management of Waiting Time u The Sof-Optics Case 1

S. D. Deshmukh OM Matching Supply and Demand u Goods vs. Services –Make to Stock vs. Make to Order –Produce in advance vs. on demand –Safety Inventory vs. Safety Capacity u Examples –Banks (tellers, ATMs, drive-ins) –Fast food restaurants (counters, drive-ins) –Retail (checkout counters) –Airline (reservation, check-in, takeoff, landing, baggage claim) –Hospitals (ER, OR, HMO) –Call centers (telemarketing, help desks, 911 emergency) –Service facilities (repair, job shop, ships/trucks load/unload) 2

S. D. Deshmukh OM Sales Reps Processing Calls (Service Process) Incoming Calls (Customer Arrivals) Calls on Hold (Service Inventory) Answered Calls (Customer Departures) Blocked Calls (Due to busy signal) Abandoned Calls (Due to long waits) The DesiTalk Call Center Calls In Process (Due to long waits) The Call Center Process 3

S. D. Deshmukh OM The Service Process u Customer Inflow (Arrival) Rate (R i ) –Inter-arrival Time = 1 / R i u Processing Time T p –Processing Rate per Server = 1/ T p u Number of Servers (c) –Number of customers that can be processed simultaneously u Total Processing Rate (Capacity) = R p = c / T p u Buffer Capacity (K) –Maximum Queue Length

S. D. Deshmukh OM Operational Performance Measures u Flow time T=T i +T p u Inventory I= I i + I p  Flow Rate R =Min (R i, R p  u Stable Process= R i < R p,, so that R = R i  Little’s Law: I = R i  T, I i = R i  T i, I p = R i  T p  Capacity Utilization  = I p / c = R i  T p / c = R i / R p < 1 u Safety Capacity R s = R p - R i  Number of Busy Servers = I p = c  = R i  T p u Fraction Lost P b = P(Blocking) = P(Queue = K)

S. D. Deshmukh OM Financial Performance Measures u Sales –Throughput Rate –Abandonment Rate –Blocking Rate u Cost –Capacity utilization –Number in queue / in system u Customer service –Waiting Time in queue /in system 6

S. D. Deshmukh OM Flow Times with Arrival Every 4 Secs Customer Number Arrival Time Departure Time Time in Process What is the queue size? What is the capacity utilization? 7

S. D. Deshmukh OM Customer Number Arrival Time Departure Time Time in Process Flow Times with Arrival Every 6 Secs What is the queue size? What is the capacity utilization? 8

S. D. Deshmukh OM Customer Number Arrival Time Processing Time Time in Process Effect of Variability What is the queue size? What is the capacity utilization? 9

S. D. Deshmukh OM Customer Number Arrival Time Processing Time Time in Process Effect of Synchronization What is the queue size? What is the capacity utilization? 10

S. D. Deshmukh OM Conclusion u If inter-arrival and processing times are constant, queues will build up if and only if the arrival rate is greater than the processing rate u If there is (unsynchronized) variability in inter- arrival and/or processing times, queues will build up even if the average arrival rate is less than the average processing rate u If variability in interarrival and processing times can be synchronized (correlated), queues and waiting times will be reduced 11

S. D. Deshmukh OM Summary: Causes of Delays and Queues u High Unsynchronized Variability in –Interarrival Times –Processing Times  High Capacity Utilization  = R i / R p, or Low Safety Capacity R s = R p – R i, due to –High Inflow Rate R i –Low Processing Rate R p = c/ T p 12

S. D. Deshmukh OM The Queue Length Formula Utilization effect Variability effect x where  R i / R p, where R p = c / T p, and C i and C p are the Coefficients of Variation (Standard Deviation/Mean) of the inter-arrival and processing times (assumed independent)

S. D. Deshmukh OM Variability Increases Average Flow Time T Utilization (ρ)  100% TpTp Throughput- Delay Curve 14

S. D. Deshmukh OM Computing Performance Measures u Given –Interarrival times: 10, 10, 2, 10, 1, 3, 7, 9, and 2 –Processing times: 7, 1, 7, 2, 8, 7, 4, 8, 5, 1 –c = 1 u Compute –Capacity Utilization  = R i / R p = –C i = 3.937/6 = –C p = /5 = u Queue Length Formula –I i = u Hence –T i = I i / R = 9.38 seconds, and T p = 5 seconds, so –T = seconds, so –I = RT = 14.38/6 =

S. D. Deshmukh OM Effect of Increasing Capacity u Given –Interarrival times: 10, 10, 2, 10, 1, 3, 7, 9, and 2 –Processing times: 7, 1, 7, 2, 8, 7, 4, 8, 5, 1 –c = 2 u Compute –Capacity Utilization  = R i / R p = –C i = 3.937/6 = –C p = /5 = u Queue Length Formula –I i = u Hence –T i = I i / R = seconds, and T p = 5 seconds, so –T = seconds, so –I = RT = /6 =

S. D. Deshmukh OM The Exponential Model u Poisson Arrivals –Infinite pool of potential arrivals, who arrive completely randomly, and independently of one another, at an average rate R i  constant over time u Exponential Processing Time –Completely random, unpredictable, i.e., during processing, the time remaining does not depend on the time elapsed, and has mean T p u Computations –C i = C p = 1 –If c = 1, T = 1/(R p  R i ), then I = R i  T,... –If c ≥ 2, and K < ∞, use Performance.xls 17

S. D. Deshmukh OM Example u Interarrival time = 6 secs, so R i = 10/min u T p = 5 secs = mins cρRsRs IiIi TiTi TI

S. D. Deshmukh OM Synchronization u Matching Capacity with Demand u Capacity –Short term Control –Long term Planning u Demand –Pricing –Scheduling 19

S. D. Deshmukh OM Performance Improvement Levers u Capacity Utilization / Safety Capacity –Demand Management (arrival rate) »Peak load pricing –Increase Capacity (processing rate) »Number of Servers (scale) »Processing Rate (speed) u Variability Reduction –Arrival times »Scheduling, Reservations, Appointments –Processing times »Standardization, Specialization, Training u Synchroniztion –Matching capacity with demand 20

S. D. Deshmukh OM Server 1 Queue 1 Server 2 Queue 2 Server 1 Queue Server 2 Effect of Pooling RiRi RiRi R i /2 21

S. D. Deshmukh OM Effect of Buffer Capacity u Process Data –R i = 20/hour, T p = 2.5 mins, c = 1, K = # Lines – c u Performance Measures K456 IiIi TiTi PbPb R 

S. D. Deshmukh OM Economics of Capacity Decisions u Cost of Lost Business C b –$ / customer –Increases with competition u Cost of Buffer Capacity C k –$/unit/unit time u Cost of Waiting C w –$ /customer/unit time –Increases with competition u Cost of Processing C s –$ /server/unit time –Increases with 1/ T p u Tradeoff: Choose c, T p, K –Minimize Total Cost/unit time = C b R i P b + C k K + C w I (or I i ) + c C s

S. D. Deshmukh OM Optimal Buffer Capacity u Cost Data –Cost of telephone line = $5/hour, Cost of server = $20/hour, Margin lost = $100/call, Waiting cost = $2/customer/minute u Effect of Buffer Capacity on Total Cost K$5(K + c)$20 c$100 R i P b $120 I i TC ($/hr)

S. D. Deshmukh OM Optimal Processing Capacity cK = 6 – cPbPb IiIi TC ($/hr) = $20c + $5(K+c) + $100R i P b + $120 I i $ $ $ $

S. D. Deshmukh OM Performance Variability u Effect of Variability –Average versus Actual Flow time u Time Guarantee –Promise u Service Level –P(Actual Time  Time Guarantee) u Safety Time –Time Guarantee – Average Time u Probability Distribution of Actual Flow Time –P(Actual Time  t) = 1 – EXP(- t / T) 26

S. D. Deshmukh OM Flow Time Management: Review u Waiting occurs due to –low processing capacity in relation to the inflow rate –variability in inter-arrival and processing times u Waiting can be reduced by –managing demand –pooling arrival streams –increasing capacity (number of servers  service rate) –reducing the variability in arrivals and processing u Optimal level of service involves a tradeoff –cost of waiting, lost business and cost of service 27

S. D. Deshmukh OM Flow Time Management Levers u Manage Arrivals –Demand Management: Price incentives –Pool arrivals u Increase Capacity –Scale: Servers, Part-timers, customer participation –Speed: Simplify, Automation, Information, Training u Decrease Variability –Arrivals: Forecast, Reservations, Pooling –Processing: Standardize u Reduce Impact of Waiting –Comfortable, Distract, Entertain, Perception 28