考虑顾客耐心的呼叫中心人力资源配置模型

合集下载
  1. 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
  2. 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
  3. 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。

[8,9]
. Ibrahim
[11]
, ,
, .
2
2.1
, s ( ,n ),
M/M/N+M. , 1 , .
λ (n 0) . θ
O (1 D ) p b (n)
Ё䗨ᬒᓗ T '
µ
T,
OD
,
1
ࠄ䖒˖O
Ⳉ᥹䗔ߎ
O 1 D
ㄝᕙᦤ冫Ⳉ ᥹䗔ߎ
OD '
ᥦ䯳Ҏ᭄n
2
Ă
S
ᥦ䯳ֵᙃ
1 Fig. 1 Schematic representation of call center with delay information
, .
,W , [15] . ψn+1 “
,
W |S
” , ,
n+1
.
E [W |S ]
n+1
,
ψn+1 =
i=1
1−
θ , sµ + iθ ′

(18)
,
1 E [W |S ] = S (1 − α0 ) 1 − pb (n) ψn+1 p(s + n)E[Wn+1 ]. p n=0
2.3

(19)
1
, . ,
: 2013−07−05; : : 2013−09−05. (71271052; 71021061).
[1,2]
, . .
5
:
687
,
[3,4] [5,6]
, , Guo . ,
[7]
.
, , , , . , Armony
[10]
. Jouini (balking) ; , . , M/M/N+M, (reneging)
28 2013
5 10 JOURNAL OF SYSTEMS ENGINEERING
Vol.28 No.5 Oct. 2013
,
(
: M/M/N+M . , . . : ; : N945; O2 24 ; ; :A , , ,
,
,
,
. . ( , . ; : 1000−5781(2013)05−0686−08 , ) ,
,
. (20)
Min s cb pB
, cr
pR
s.t. cb pB + cr pR
, cb
γ,
,
cr
690
28
, , , cb < cr . γ , , “
. ”
, .
3
s,
, “ ”
R
.
, (21)
θ,

λ = θ ′ L.
(13), (17) (21),

θ′ =
, L, p(s + n), rn (ϕ)
(10)
Βιβλιοθήκη Baidu
ϕ
U (a, b) 1 rn (ϕ) = b−a
b a b a
1−β −
i=0 n n
sµ + jθ ′ (j − i)θ ′ j =0,j =i
′ sµ + iθ ′ e−(sµ+iθ )wn ′ sµ + θ/ϕ + iθ

=1−β −
i=0
sµ + jθ ′ (j − i)θ ′ j =0,j =i
110819)
Staffing model for call center with customer patience variation
Yu Miao, Gong Jun, Luo Xinggang, Zhu Huabo
(Department of Systems Engineering, Key Lab of Integrated Automation of Process Industry of MOE, Northeastern University, Shenyang 110819, China)
n n
(sµ + iθ ′ ) e−(sµ+iθ )t , t 0,
Gn (t) = 1 −
i=0 n
sµ + jθ ′ −(sµ+iθ′ )t e ,t ′ ( j − i ) θ j =0,j =i
(6) (7) ,
E[Wn ] =
i=1
1 . sµ + iθ ′
3
rn (ϕ)
n+1 rn (ϕ) = P {T ′ < Wn |T > wn )} .
λ L
α′ p(s + n)rn (ϕ).
n=0
(22)
pb (n)
f (θ ′ ), ,
,
, θ′
θ′ θ′
, . .
(22) (22) . , . , , 2.2
f (θ ′ ), f
θ′ =
f θ

. ,

θ
s
wn , r (ϕ), λR , L, , p(i). , .
1 2
. , Gn (t) t wn (i) − wn (i − 1) < 10−4 , i wn (i − 1) , , θ′, :
θ′
(4)
,
T ′ = ϕT + (1 − ϕ)wn ,
688
28
T
. (4) , [9] ,
, θ′
“ ,

[12]
,ϕ ;
ϕ=1
.
T
,
wn
ϕ < 0. ϕ
,
ϕ=0 , ,
U (a, b). Wn
, ,
s+n+1 s , ′ ′ ′ sµ + θ , sµ + 2θ , . . . , sµ + nθ n
′ sµ + iθ ′ e−(sµ+iθ )wn × ′ sµ + θ/ϕ + iθ
1−
2.2
θ ln (b − a)(sµ + iθ ′ )
3 , ,
(b − a)(sµ + iθ ′ ) +1 θ
, ,
(11)
5
:
689
.
[13]
L(t)
,
3
O
0 1 Ă
S-1
t (t ,
O
S
0)
,
O(1D)(1 pb (1))
,
5
:
691
. 1 ,β . , .
1 β (cb = 1, cr = 1) Table 1 Staffing and performance measures for different β with cb = 1, cr = 1 β 10% 30% 50% 70% 90% pB 0.057 0 0.093 9 0.158 1 0.216 0 0.277 2 pR 0.208 0 0.168 8 0.137 9 0.074 7 0.017 2 pS 0.735 0 0.737 3 0.704 0 0.709 3 0.705 7 E [W |S ] 0.019 4 0.024 0 0.041 4 0.052 6 0.052 5 s 16 16 15 15 15
θ ′(0) ← θ, i ← 0, ε; θ
′(i)
−θ θ ′(i−1)
′(i−1)
>ε λ
R(i)
,

θ ′ = f (θ ′ )
,
=
n=0 ∞
λα′ (θ ′(i−1) ))p(s + n)(θ ′(i−1) )rn (θ ′(i−1) ), ip(s + i)(θ ′(i−1) );
L(i) =
S+1 S+2
, {L(t), t
0}
.
s
O(1D)(1 pb(0))
P
Fig. 3
sP
3
sP T '
sP 2T '
The birth-and-death process of the model
p(i)
i i λ i!µi p(0), p(i) = λi (1 − α)i−s s!µs
Abstract: The impact of impatience psychology on customer behaviors plays an important role in the human resources staffing of call centers. The paper formulates a model of staffing optimization of call centers considering stochastic customer patience variation on staffing, based on M/M/N+M queue with delay information announced. Firstly, the queuing model with delay information is given based on queuing theory, and customer patience variation and the probability of customers’ behavior of abandonment including balking and reneging are formulated as a function of the delay information. Secondly, we derive some related performances by analyzing the queuing process for steady state probability. Lastly, the optimal staffing level is calculated through a bisection algorithm and the fixed point algorithm. The numerical example illustrates that how the reliability probability of delay information, penalty coefficient of customer behavior, and patience variation affect staffing. The conclusion drawn has guiding implications for human resource management in call centers. Key words: call center; customer patience; delay information; human resources staffing; balking and reneging
i−s
1 1 − pb (j − 1) sµ + jθ ′ j =1 1 − pb (j − 1) sµ + jθ ′ j =1 p ,
S i−s i−s
i
s
(12)
p(0), i > s,
−1
p(0) =
∞ λi (1 − α) λi + i s!µs i=0 i!µ i=s+1 s
.
PASTA(Poison arrivals see time averages)[14] ,
s
sP
Wn s+n+1 s
,
Ă
s+n
s P +nT '
, “
2 ”. .
. , Wn
Wn
s+1
s+n+1
2 Fig. 2 The diagram of the pure-death process
gn (t)
Wn
n
, Gn (t)
n
Wn

, E[Wn ]

Wn 0,
(5)
gn (t) =
i=0
sµ + jθ (j − i)θ ′ j =0,j =i
(8)
(1)
(8)
∞ wn
rn (ϕ) = P {ϕT + (1 − ϕ)wn < Wn |T > wn )} = 1 − β − eθ(wn /ϕ)
,
n
gn (t)e−θt/ϕ dt.
(9)
rn (ϕ) = 1 − β −
i=0
sµ + jθ ′ (j − i)θ ′ j =0,j =i
n n
n
′ sµ + iθ ′ e−(sµ+iθ )wn , ′ sµ + θ/ϕ + iθ
Wn wn
1
,
:
β wn
,
,
Wn
(1) ; ( , ), (2)
β = P {Wn < wn } .
, , , , “
s, α;
”,
pb (n) = P {T < wn } = 1 − e−θwn ,
,
α′ α′ = (1 − α)(1 − pb (n)).
, (3)
[9]
T T′
,
, . T′
.
2
i=1
3 4
θ ′(i) , +1,
θ ′(i) = λR(i) /L(i) ;
2.
4
, . 2GB ,λ
4.1
. , .
β, MATLAB , CPU Intel Core 2 (2.67 GHz), µ = 1 ,α = 0.05, γ = 0.3 , θ = 0.5 . ,
,
λ = 20, a = 0, b = 1 , cb = cr = 1, β
, pB
, ,
, pR
,
s
,
β β β
,s .
s
,
p E [W |S ]
S
p . β
S
s
,
4.2
λ = 20, a = 0, b = 1 ,
, 1
2
cb = 1, ,
cr ,
1, 2
3.
cb /cr
, .
γ
p
R
2
3
, ,
.
β (cb = 1, cr = 3) Table 3 Staffing and performance measures for different β with cb = 1, cr = 3 β 10% 30% 50% 70% 90% pB 0.024 0 0.039 7 0.070 8 0.138 9 0.242 2 pR 0.086 6 0.068 5 0.059 7 0.047 2 0.014 9 pS 0.889 3 0.891 9 0.869 5 0.813 8 0.742 9 E [W |S ] 0.006 0 0.007 8 0.013 2 0.026 4 0.041 6 s 21 21 20 18 16 3
p ,
B
p ,

R
λ ,
R
L, (13)– (17)
. (13) (14) (15) (16) (17)
L=
i=1 ∞
ip(s + i), (α + (1 − α)pb (n))p(n + s),
n=0 ′
pB = pR =
θL , λ pS = 1 − pR − pB ,

λR =
n=0
λα′ p(s + n)rn (ϕ).
相关文档
最新文档