A best possible algorithm for an online scheduling problem with position-based learning effect

Ran Ma; Lu Zhang; Yuzhong Zhang

doi:10.3934/jimo.2021144

Article Contents

2022, Volume 18, Issue 6: 3989-4010. Doi: 10.3934/jimo.2021144

This issue Previous Article Nonmonotone spectral gradient method based on memoryless symmetric rank-one update for large-scale unconstrained optimization Next Article Optimal per-loss reinsurance and investment to minimize the probability of drawdown

A best possible algorithm for an online scheduling problem with position-based learning effect

1.
School of Management Engineering, , Qingdao University of Technology, Qingdao 266525, China, University Research Center for Smart City Construction and Management of Shandong Province, Qingdao 266525, China
2.
Institute of Operations Research, School of Management, Qufu Normal University, Rizhao 276826, China

^*Corresponding author: Ran Ma
^*Corresponding author: Ran Ma

Received: January 2021

Revised: April 2021

Early access: September 2021

Published: November 2022

This research is supported by the National Natural Science Foundation of China (Grant Nos. 11501171, 11771251) and the Province Natural Science Foundation of Shandong (Grant No. ZR2020MA028).

Abstract / Introduction Full Text(HTML) Figure(3) / Table(14) Related Papers Cited by

Abstract

In this paper, we focus on an online scheduling problem with position-based learning effect on a single machine, where the jobs are released online over time and preemption is not allowed. The information about each job $ J_j $, including the basic processing time $ p_j $ and the release time $ r_j $, is only available when it arrives. The actual processing time $ p_j' $ of each job $ J_j $ is defined as a function related to its position $ r $, i.e., $ p_j' = p_j(\alpha-r\beta) $, where $ \alpha $ and $ \beta $ are both nonnegative learning index. Our goal is to minimize the sum of completion time of all jobs. For this problem, we design a deterministic polynomial time online algorithm Delayed Shortest Basic Processing Time (DSBPT). In order to facilitate the understanding of the online algorithm, we present a relatively common and simple example to describe the execution process of the algorithm, and then by competitive analysis, we show that online algorithm DSBPT is a best possible online algorithm with a competitive ratio of 2.

Keywords:

Mathematics Subject Classification: Primary: 90B35.

Citation:

Full Text(HTML)

Figure 1. The scheduling of DSBPT and an off-line optimal scheduling

Download: Full-size image PowerPoint slide

Figure 2. Block and subblock

Download: Full-size image PowerPoint slide

Figure 3. Reverse pair

Download: Full-size image PowerPoint slide

Table 1. Notations-1

Notation	Meaning
$ J_j $	the job of index $ j $, where $ j=1,2,\ldots,n $
$ I $	a job instance, $ I=\left\{J_1,J_2,\ldots,J_n\right\} $
$ r_j(I) $	the release time of job $ J_j $ in $ I $
$ p_j(I) $	the basic processing time of job $ J_j $ in $ I $
$ p'_j(I) $	the actual processing time of job $ J_j $ in $ I $
$ X(I) $	a feasible schedule of $ I $
$ S_j(X(I)) $	the starting time of job $ J_j $ in $ X(I) $
$ C_j(X(I)) $	the completion time of job $ J_j $ in $ X(I) $,
	$ C_j(X(I))=S_j(X(I))+p_j(I) $
$ Z_j(X(I)) $	the contribution of job $ J_j $ to the total objective value of $ I $ in $ X(I) $
$ Z(X(I)) $	the total objective value of $ I $ in $ X(I) $,
	$ Z(X(I))=\sum Z_j(X(I)) $
$ \pi(I) $	an off-line optimal schedule of $ I $
$ {\rm{OPT}}( I )$	the total objective value of $ I $ in $ \pi(I) $, $ {\rm{OPT}}(I)=\sum Z_j(\pi(I)) $

| Show Table

DownLoad: CSV

Table 2. Notations-2

Notation	Meaning
$ \|T\| $	the number of jobs in $ T $
$ T(X(I)) $	the set of jobs of $ T $ in $ X(I) $
$ Z_T(X(I)) $	the total contribution of jobs of $ T $ to $ Z(X(I)) $,
	$ Z_T(X(I))=\sum_{J_j\in T} Z_j(X(I)) $
$ {\rm{OPT}}(T) $	the total objective value of jobs of $ T $ in an off-line optimal schedule, $ {\rm{OPT}}(T)=\sum Z_j(\pi(T)) $

| Show Table

DownLoad: CSV

Algorithm 1 DSBPT

At each decision time $ t $, if the machine is idle and there exist valid jobs, the valid job with the shortest basic processing time is selected. If there are many such valid jobs, the job $ J_j $ with the earliest release time is selected, presume its processing position is $ k $ if $ p_j(I)(\alpha-k\beta)\leq t $, and process the job; otherwise the machine will remain idle and wait until the next decision time $ t $ to make a decision.

| Show Table

DownLoad: CSV

Algorithm 1: Online Algorithm DSBPT.

1: Input: job instance $ I $ ($ |I|\geq 1 $) and learning index $ \alpha $ and $ \beta $
2: $ S $ is an empty sequence, $ k\leftarrow 1 $, $ C\leftarrow 0 $, $ A $ is an available jobs set, $ f(S)\leftarrow 0 $, $ dt\leftarrow 0 $, $ Ct\leftarrow 0 $, $ t\leftarrow \min\{r_j|J_j\in I\} $
3: While True do
4:       update available jobs set A, $ dt\leftarrow \max\{t,Ct\} $
5:       $ D \leftarrow \{J_j\in A|\ the \ job \ has \ the \ shortest \ processing \ time \ first $
            $ \qquad \qquad \qquad \qquad \ and \ then \ releases \ the \ earliest\} $
6:       arbitrarily select a job $ J_a $ from $ D $
7:       If $ dt \geq p_a(\alpha-k\beta) $ then
8:          $ S\leftarrow S + \{J_a\} $
9:          $ C\leftarrow dt+p_a(\alpha-k\beta) $
10:          $ f(S)\leftarrow f(S)+C $
11:          $ k\leftarrow k+1 $
12:       end if
13:       If $ C>dt $ then
14:             $ Ct\leftarrow C $
15:       end if
16: end while
17: return $ S $ and $ f(S) $

| Show Table

DownLoad: CSV

Table 3. Information of instance $ I $

	$ J_1 $	$ J_2 $	$ J_3 $	$ J_4 $	$ J_5 $	$ J_6 $
release time	0	8	17	23	28	30
basic processing time	8	6	12	10	4	8

| Show Table

DownLoad: CSV

Table 4. The selection process of jobs under $ \sigma(I) $

decision time	valid job set	selected job	sequence of finished jobs	current objective function value
0	$ J_1 $	None	None	0
8	$ J_1 $ $ J_2 $	$ J_2 $	None	0
14	$ J_1 $	$ J_1 $	$ J_2 $	14
17	$ J_3 $	None	$ J_2 $	14
20.8	$ J_3 $	$ J_3 $	$ J_2 $-$ J_1 $	34.8
23	$ J_4 $	None	$ J_2 $-$ J_1 $	34.8
28	$ J_4 $ $ J_5 $	None	$ J_2 $-$ J_1 $	34.8
29.2	$ J_4 $ $ J_5 $	$ J_5 $	$ J_2 $-$ J_1 $-$ J_3 $	64
30	$ J_4 $ $ J_6 $	None	$ J_2 $-$ J_1 $-$ J_3 $	64
31.4	$ J_4 $ $ J_6 $	$ J_6 $	$ J_2 $-$ J_1 $-$ J_3 $-$ J_5 $	95.4
34.6	$ J_4 $	$ J_4 $	$ J_2 $-$ J_1 $-$ J_3 $-$ J_5 $-$ J_6 $	130
37.1	None	None	$ J_2 $-$ J_1 $-$ J_3 $-$ J_5 $-$ J_6 $-$ J_4 $	167.1

| Show Table

DownLoad: CSV

Table 5. The selection process of jobs under $ \pi(I) $

time	valid job set	selected job	sequence of finished jobs	current objective function value
0	$ J_1 $	$ J_1 $	None	0
8	$ J_2 $	$ J_2 $	$ J_1 $	8
13.1	None	None	$ J_1 $-$ J_2 $	21.1
17	$ J_3 $	$ J_3 $	$ J_1 $-$ J_2 $	21.1
23	$ J_4 $	None	$ J_1 $-$ J_2 $	21.1
25.4	$ J_4 $	$ J_4 $	$ J_1 $-$ J_2 $-$ J_3 $	46.5
28	$ J_5 $	None	$ J_1 $-$ J_2 $-$ J_3 $	46.5
30	$ J_5 $ $ J_6 $	None	$ J_1 $-$ J_2 $-$ J_3 $	46.5
30.9	$ J_5 $ $ J_6 $	$ J_5 $	$ J_1 $-$ J_2 $-$ J_3 $-$ J_4 $	77.4
32.5	$ J_6 $	$ J_6 $	$ J_1 $-$ J_2 $-$ J_3 $-$ J_4 $-$ J_5 $	109.9
34.5	None	None	$ J_1 $-$ J_2 $-$ J_3 $-$ J_4 $-$ J_5 $-$ J_6 $	144.4

| Show Table

DownLoad: CSV

Table 6. The comparison between $ \sigma(I) $ and $ \pi(I) $

	job processing scheduling	objective function value
$ \sigma(I) $	$ J_2-J_1-J_3-J_5-J_6-J_4 $	167.1
$ \pi(I) $	$ J_1-J_2-J_3-J_4-J_5-J_6 $	144.4

| Show Table

DownLoad: CSV

Algorithm 2 FDSBPT

At arbitrary time $ t $, if the machine is idle and there exist valid jobs, the valid job with the shortest basic processing time is selected. If there are multiple jobs with the shortest processing time, choose arbitrary job $ J_j $ flexibly, presume its processing position is k, if $ p_j(I)(\alpha-k\beta)\leq t $, we arrange job $ J_j $ on the idle machine at time $ t $; otherwise, we do nothing until the next time.

| Show Table

DownLoad: CSV

Algorithm 2: Online Algorithm DSBPT.

1: Input: job instance $ I $ ($ |I|\geq 1 $) and learning index $ \alpha $ and $ \beta $
2: $ S $ is an empty sequence, $ k\leftarrow 1 $, $ C\leftarrow 0 $, $ A $ is an available jobs set, $ f(S)\leftarrow 0 $, $ t\leftarrow 0 $
3: While True do
4:       update available jobs set A
5:       $ D \leftarrow \{ {J_j} \in A|\;the\;job\;has\;the\;shortest\;processing\;time{\rm{\} }} $
6:       arbitrarily select a job $ J_a $ from $ D $
7:       If $ t \geq p_a(\alpha-k\beta) $ then
8:          $ S\leftarrow S + \{J_a\} $
9:          $ C\leftarrow t+p_a(\alpha-k\beta) $
10:          $ f(S)\leftarrow f(S)+C $
11:          $ k\leftarrow k+1 $
12:       end if
13:       $ t\leftarrow t+1$
14: end while
15: return $ S $ and $ f(S) $

| Show Table

DownLoad: CSV

Table 7. Information of instance $ I_c $

	$ J_1 $	$ J_2 $	$ J_3 $	$ J_4 $	$ J_5 $	$ J_6 $	$ J_7 $	$ J_8 $	$ J_9 $	$ J_{10} $
release time	0	8	17	23	28	30	30	33	35	39
basic processing time	8	6	12	10	4	8	12	9	10	11
	$ J_{11} $	$ J_{12} $	$ J_{13} $	$ J_{14} $	$ J_{15} $	$ J_{16} $	$ J_{17} $	$ J_{18} $	$ J_{19} $	$ J_{20} $
release time	41	42	46	48	48	48	48	50	52	52
basic processing time	13	6	7	9	8	8	9	5	13	8

| Show Table

DownLoad: CSV

Table 8. The selection process of jobs under $\sigma(I_c)$

decision time	valid job set	selected job	current objective function value
0	$J_1$	None	0
8	$J_1$ $J_2$	$J_2$	0
14	$J_1$	$J_1$	14
17	$J_3$	None	14
21.84	$J_3$	$J_3$	35.84
23	$J_4$	None	35.84
28	$J_4$ $J_5$	None	35.84
30	$J_4$ $J_5$ $J_6$ $J_7$	None	35.84
33	$J_4$ $J_5$ $J_6$ $J_7$ $J_8$	None	35.84
33.36	$J_4$ $J_5$ $J_6$ $J_7$ $J_8$	$J_5$	69.2
35	$J_4$ $J_6$ $J_7$ $J_8$ $J_9$	None	69.2
37.12	$J_4$ $J_6$ $J_7$ $J_8$ $J_9$	$J_6$	106.32
39	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$	None	106.32
41	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$	None	106.32
42	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{12}$	None	106.32
44.48	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{12}$	$J_{12}$	150.8
46	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{13}$	None	150.8
48	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{13}$ $J_{14}$ $J_{15}$ $J_{16}$ $J_{17}$	None	150.8
49.88	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{13}$ $J_{14}$ $J_{15}$ $J_{16}$ $J_{17}$	$J_{13}$	200.68
50	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{15}$ $J_{16}$ $J_{17}$ $J_{18}$	None	200.68
52	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{15}$ $J_{16}$ $J_{17}$ $J_{18}$ $J_{19}$ $J_{20}$	None	200.68
56.04	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{15}$ $J_{16}$ $J_{17}$ $J_{18}$ $J_{19}$ $J_{20}$	$J_{18}$	256.72
60.34	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{15}$ $J_{16}$ $J_{17}$ $J_{19}$ $J_{20}$	$J_{15}$	317.06
67.06	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{16}$ $J_{17}$ $J_{19}$ $J_{20}$	$J_{16}$	384.12
73.62	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{17}$ $J_{19}$ $J_{20}$	$J_{20}$	457.74
80.02	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{17}$ $J_{19}$	$J_8$	537.76
87.04	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{17}$ $J_{19}$	$J_{14}$	624.8
93.88	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{17}$ $J_{19}$	$J_{17}$	718.68
100.54	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{19}$	$J_4$	819.22
107.74	$J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{19}$	$J_9$	926.96
114.74	$J_7$ $J_{10}$ $J_{11}$ $J_{19}$	$J_{10}$	1041.7
122.22	$J_7$ $J_{11}$ $J_{19}$	$J_7$	1163.92
130.14	$J_{11}$ $J_{19}$	$J_{11}$	1294.06
138.46	$J_{19}$	$J_{19}$	1432.52
146.52	None	None	1579.04

| Show Table

DownLoad: CSV

Table 9. The selection process of jobs under π(I_c)

time	valid job set	selected job	current objective function value
0	$J_1$	$J_1$	0
8	$J_2$	$J_2$	8
13.88	None	None	21.88
17	$J_3$	$J_3$	21.88
28.52	$J_4$ $J_5$	$J_5$	50.4
32.28	$J_4$ $J_6$ $J_7$	$J_6$	82.68
39.64	$J_4$ $J_7$ $J_8$ $J_9$ $J_{10}$	$J_8$	122.32
47.74	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{12}$ $J_{13}$	$J_{12}$	170.06
53.02	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{13}$ $J_{14}$ $J_{15}$ $J_{16}$ $J_{17}$ $J_{18}$ $J_{19}$ $J_{20}$	$J_{18}$	223.08
57.32	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{13}$ $J_{14}$ $J_{15}$ $J_{16}$ $J_{17}$ $J_{19}$ $J_{20}$	$J_{13}$	280.4
63.2	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{15}$ $J_{16}$ $J_{17}$ $J_{19}$ $J_{20}$	$J_{15}$	343.6
69.76	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{16}$ $J_{17}$ $J_{19}$ $J_{20}$	$J_{16}$	413.36
76.16	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{17}$ $J_{19}$ $J_{20}$	$J_{20}$	489.52
82.4	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{14}$ $J_{17}$ $J_{19}$	$J_{14}$	571.92
89.24	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{17}$ $J_{19}$	$J_{17}$	661.16
95.9	$J_4$ $J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{19}$	$J_4$	757.06
103.1	$J_7$ $J_9$ $J_{10}$ $J_{11}$ $J_{19}$	$J_9$	860.16
110.1	$J_7$ $J_{10}$ $J_{11}$ $J_{19}$	$J_{10}$	970.26
117.58	$J_7$ $J_{11}$ $J_{19}$	$J_7$	1087.84
125.5	$J_{11}$ $J_{19}$	$J_{11}$	1213.34
133.82	$J_{19}$	$J_{19}$	1347.16
141.88	None	None	1489.04

| Show Table

DownLoad: CSV

Table 10. The comparison between $ \sigma(I_c) $ and $ \pi(I_c) $

	job processing scheduling	objective function value
$ \sigma(I_c) $	$ J_2-J_1-J_3-J_5-J_6-J_{12}-J{13}-J_{18}-J_{15}-J_{16}-J_{20}-J_8-J_{14}-J_{17}-J_4-J_9-J_{10}-J_7-J_{11}-J_{19} $	1579.04
$ \pi(I_c) $	$ J_1-J_2-J_3-J_5-J_6-J_8-J_{12}-J_{18}-J{13}-J_{15}-J_{16}-J_{20}-J_{14}-J_{17}-J_4-J_9-J_{10}-J_7-J_{11}-J_{19} $	1489.04

| Show Table

DownLoad: CSV

Related Papers

Cited by

References

[1]	W. Allihaibi, M. Cholette, M. Masoud, J. Burke and A. Karim, A heuristic approach for scheduling patient treatment in an emergency department based on bed blocking, International Journal of Industrial Engineering Computations, 11 (2020), 565-584. doi: 10.5267/j.ijiec.2020.4.005.
[2]	D. Y. Bai, H. Y. Xue, L. Wang, C. C. Wu, W. C. Lin and D. H. Abdulkadir, Effective algorithms for single-machine learning-effect scheduling to minimize completion-time-based criteria with release dates, Expert Systems With Applications, 156 (2020), 113445.
[3]	L. Bai, D. Yang, X. Wang, L. Tong, X. Zhu, N. Zhong, C. Bai and C. A. Powell, Chinese experts consensus on the Internet of Things-aided diagnosis and treatment of coronavirus disease 2019 (COVID-19), Clinical eHealth, 3 (2020), 7-15. doi: 10.1016/j.ceh.2020.03.001.
[4]	D. Biskup, Single-machine scheduling with learning considerations, European J. Oper. Res., 115 (1999), 173-178. doi: 10.1016/S0377-2217(98)00246-X.
[5]	D. Biskup, A state-of-the-art review on scheduling with learning effects, European J. Oper. Res., 188 (2008), 315-329. doi: 10.1016/j.ejor.2007.05.040.
[6]	T. C. E. Cheng, S. R. Cheng, W. H. Wu, P. H. Hsu and C. C. Wu, A two-agent single machine scheduling problem with truncated sum-of-processing-times-based learning considerations, Computers and Industrial Engineering, 60 (2011), 534-541.
[7]	M. B. Cheng, S. X. Xiao and G. Liu, Single-machine rescheduling problems with learning effect under disruptions, J. Ind. Manag. Optim., 14 (2018), 967-980. doi: 10.3934/jimo.2017085.
[8]	J. A. Hoogeveen and A. P. A. Vestjens, Optimal on-line algorithms for single-machine scheduling, Lecture Notes in Comput. Sci., Springer, Berlin, (1996), 404–414.
[9]	Z. Y. Jiang, F. F. Chen and X. D. Zhang, Single-machine scheduling with times-based and job-dependent learning effect, Journal of the Operational Research Society, 68 (2017), 809-815.
[10]	S. Jun and S. Lee, Learning dispatching rules for single machine scheduling with dynamic arrivals based on decision trees and feature construction, International Journal of Production Research, 59 (2020), 2838-2856. doi: 10.1080/00207543.2020.1741716.
[11]	E. L. Lawler, J. K. Lenstra, A. H. G. Rinnooy and D. B. Shmoys, Sequencing and scheduling algorithms and complexity, Handbooks in Operations Research and Management Science, 4 (1993), 445-522.
[12]	W. C. Lee and C. C. Wu, A note on single-machine group scheduling problems with position-based learning effect, Appl. Math. Model., 33 (2009), 2159-2163. doi: 10.1016/j.apm.2008.05.020.
[13]	P. H. Liu and X. W. Lu, On-line scheduling of parallel machines to minimize total completion times, Comput. Oper. Res., 36 (2009), 2647-2652. doi: 10.1016/j.cor.2008.11.008.
[14]	Y. Y. Lu, F. Teng and Z. X. Feng, Scheduling jobs with truncated exponential sum-of-logarithm-processing-times based and position-based learning effects, Asia-Pac. J. Oper. Res., 32 (2015), 1550026. doi: 10.1142/S0217595915500268.
[15]	R. Ma and J. P. Tao, An improved 2.11-competitive algorithm for online scheduling on parallel machines to minimize total weighted completion time, J. Ind. Manag. Optim., 14 (2018), 497-510. doi: 10.3934/jimo.2017057.
[16]	R. Ma and S. N. Guo, Applying "Peeling Onion" approach for competitive analysis in online scheduling with rejection, European J. Oper. Res., 290 (2021), 57-67. doi: 10.1016/j.ejor.2020.08.009.
[17]	G. Mosheiov, Scheduling problems with a learning effect, European J. Oper. Res., 132 (2001), 687-693. doi: 10.1016/S0377-2217(00)00175-2.
[18]	G. Mosheiov, Minimizing total absolute deviation of job completion times: Extensions to position-dependent processing times and parallel identical machines, J. Oper. Res. Soc., 59 (2008), 1422-1424. doi: 10.1057/palgrave.jors.2602480.
[19]	S. Mustu and T. Eren, The single machine scheduling problem with setup times under an extension of the general learning and forgetting effects, Optim. Lett., 15 (2021), 1327-1343. doi: 10.1007/s11590-020-01641-9.
[20]	J. B. Wang, M. Gao, J. J. Wang, L. Liu and H. Y. He, Scheduling with a position-weighted learning effect and job release dates, Eng. Optim., 52 (2019), 1475-1493. doi: 10.1080/0305215X.2019.1664498.
[21]	J.-B. Wang, L. H. Sun and L. Y. Sun, Single machine scheduling with exponential sum-of-logarithm-processing-times based learning effect, Appl. Math. Model., 34 (2010), 2813-2819. doi: 10.1016/j.apm.2009.12.015.
[22]	J.-B. Wang and J.-J. Wang, Single machine scheduling with sum-of-logarithm processing-times based and position based learning effects, Optim. Lett., 8 (2014), 971-982. doi: 10.1007/s11590-012-0494-4.
[23]	J. B. Wang and Z. Q. Xia, Flow-shop scheduling with a learning effect, J. Oper. Res. Soc., 56 (2005), 1325-1330. doi: 10.1057/palgrave.jors.2601856.
[24]	T. P. Wright, Factors affecting the cost of airplanes, J. Aer. Sci., 3 (1936), 122-128. doi: 10.2514/8.155.
[25]	S.-J. Yang and D.-L. Yang, Note on "A note on single-machine group scheduling problems with position-based learning effect", Appl. Math. Model., 34 (2010), 4306-4308. doi: 10.1016/j.apm.2010.03.037.
[26]	A. E. Zade, S. S. Haghighi and M. Soltani, Reinforcement learning for optimal scheduling of Glioblastoma treatment with Temozolomide, Computer Methods and Programs in Biomedicine, 193 (2020), 105443.