Optimization Work1

M/M/s/C/N Queueing Sensitivity Simulator

Finite-population queue with s servers and limited capacity. Drag any slider, or type a value beyond its range, and the graphs update instantly.

Reference: Dhakal, D. P., Ghimire, R. P., & Poudel, B. N. (2026). The value of the customer’s waiting time for finite source queueing system. Kathmandu University Journal of Science, Engineering and Technology, 20(1). Read the paper

Utilization factor (ρ) vs waiting-time value (β), curves for different numbers of servers

Number of servers (s) vs server utilization

Number of servers (s) vs β, at your ρ

Queue length Lq and idle servers Ss vs ρ

Distribution Pn (teal: free server, orange: waiting)

Export

Interpretation of the current results / हालको नतिजाको व्याख्या

English EN

    नेपाली NE

      The text updates automatically whenever you move a slider. / स्लाइडर चलाउँदा यो व्याख्या आफैँ परिवर्तन हुन्छ।

      Mathematical model

      State probabilities, ρ = λ/μ Pn = N!n!(N−n)! ρn P0,   0 ≤ n ≤ s
      Pn = N!(N−n)! 1s!·sn−s ρn P0,   s < n ≤ C
      Normalisation P0 = 1CΣn=0 wn,   where Pn = wn P0
      Performance measures Lq = CΣn=s+1 (n−s) Pn    Ss = s−1Σn=0 (s−n) Pn
      L = CΣn=0 n Pn    λeff = λ(N − L)    Wq = Lq/λeff
      Value of the customer’s waiting time β = − ∂Ss/∂ρ∂Lq/∂ρ
      Exact derivatives (used in this simulator) ∂Lq/∂ρ = 1ρ CΣn=s+1 (n−L)(n−s) Pn    ∂Ss/∂ρ = 1ρ s−1Σn=0 (n−L)(s−n) Pn

      β is the idle-server time saved per unit increase in queue length (Dhakal et al., 2026, after Singhal et al., 2018). As ρ grows, Lq never decreases and Ss never increases, so ∂Ss/∂ρ ≤ 0 ≤ ∂Lq/∂ρ and, with the leading minus sign, β ≥ 0 always.