BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:064ad783ae5e0563d114a2187f367eb6
DTSTAMP:20240503T042424Z
SUMMARY:Staffing to Asymptotically Stabilize Performance in Nonstationary S
 ervice Systems 
DESCRIPTION: \n\nAbstract: Queueing theory is a field driven by application
 s. But\nunfortunately\, there still remains a large gap between tractable\
 ntheoretical studies and practical applications\, such as call centers\nan
 d health care systems\, which have many realistic features (e.g.\,\ntime-v
 arying arrivals\, customer abandonment\, non-exponential\ndistributions\, 
 and complicated network structures). In response to the\nchallenge\, we st
 udy a general G_t/GI/s_t+GI queueing model\, which has\na non-stationary n
 on-Poisson arrival process (the G_t)\,\nnon-exponential service times (the
  first GI)\, and allows customer\nabandonment according to a non-exponenti
 al patience distribution (the\n+GI). To bridge the gap between mathematica
 l tractability and model\napplicability\, we develop fundamental principle
 s and optimal control\npolicies for such a general queueing model. Analyti
 c formulas are\ndeveloped to set the time-dependent number of servers in o
 rder to\nstabilize important service-level indicators\, including: mean cu
 stomer\ndelay\, probability of abandonment\, and tail probability of delay
 \n(TPoD). Taking the TPoD for example: for any delay target w > 0 and\npro
 bability target 0 < alpha < 1\, we determine appropriate\ntime-dependent s
 taffing levels (the s_t) so that the time-varying\nprobability that the wa
 iting time exceeds a maximum acceptable value w\nis stabilized at alpha at
  all times. In addition\, effective\napproximating formulas are provided f
 or other important performance\nfunctions such as the probabilities of del
 ay and abandonment\, and the\nmeans of delay and queue length. Many-server
  heavy-traffic limit\ntheorems in the efficiency-driven regime are develop
 ed to show that\n(i) the proposed staffing function achieves the goal asym
 ptotically as\nthe scale increases\, and (ii) the proposed approximating f
 ormulas for\nother performance measures are asymptotically accurate as the
  scale\nincreases. Extensive simulations show that both the staffing funct
 ions\nand the performance approximations are effective\, even for smaller\
 nsystems having an average of 3 servers.\n\nMini-bio: Yunan Liu is an assi
 stant professor at the Industrial and\nSystems Engineering Department and 
 an associate faculty member of the\nOperations Research Center of North Ca
 rolina State University. His\nresearch interests include queueing theory\,
  stochastic modeling\,\napplied probability\, simulation\, and their appli
 cations in service\nsystems including call centers\, healthcare\, manufact
 uring systems and\ntransportation systems. He received his M.S. and Ph.D. 
 in Operations\nResearch from Columbia University and B.S. in Electrical En
 gineering\nfrom Tsinghua University.\n 
DTSTART:20151116T180000Z
DTEND:20151116T190000Z
LOCATION:Gerri C. LeBow Hall\, 3220 Market Street\, 722\, Philadelphia\, PA
  19104
END:VEVENT
END:VCALENDAR