Presentation on Smart Locker Bank Design: A Scenario Based Optimization Approach given be Louis Faugère at the 12th Congrès International de Génie Industriel in Compiègnes, France (2017).

1. Smart Lockers Bank Design
A Scenario Based Optimization
Approach
Louis Faugere & Benoit Montreuil
Physical InternetCenter
School of Industrial & Systems Engineering
Georgia Institute ofTechnology
Email: Louis.Faugere@gatech.edu
Louis Faugere – 03/05/17

2. Presentation Outline
2
• Smart Lockers as Pickup & Delivery Points
• Why Smart Lockers?
• Hyperconnected City Logistics
• Challenges
• Optimization Based Design Model
• Experimentation
• Conclusions

3. Sources: dhl.com, mypopstation.com, inpost.pl
Smart Lockers as Pickup & Delivery Points
3

4. Sources: http://www.foodinsight.org/blogs/; http://www.wsbradio.com/news/
Why Smart Lockers?
4

5. Why Smart Lockers?
5
Courier, Express & Parcel (CEP) IndustryTrends
Sources: “AddingValue to Parcel Delivery”, Accenture 2015

6. Hyperconnected City Logistics
6
Montreuil, Benoit. (2016). Omnichannel Business‐to‐Consumer Logistics and Supply Chains:Towards Hyperconnected Networks and Facilities

7. Hyperconnected City Logistics
7

8. Challenges
8
• Number of smart locker banks to deploy
• Location of each smart locker bank
• Design of each smart locker bank

9. Optimization Based Design Model
9
Locker Unit
Grid Unit
Maximum Grid Locker Bank
Grid Used
Interactive
Module

10. Optimization Based Design Model
10
(2 x 2)
Small
Medium
Large
(1 x 1) (1 x 2) (1 x 3)

11. Optimization Based Design Model
11
𝐻: Horizontal dimensions of LockerTerminal in number of grid units.
𝑉:Vertical dimension of LockerTerminal in number of grid units.
𝐴ℎ𝑣𝑥𝑦: 0-1 assignment of a locker with dimensions h and v, so that its lower corner is located in
grid location (x,y).
𝑀 𝑥𝑦: 0-1 assignment of the customer interactive module so that its lower corner is located in
grid location (x,y).
𝑁ℎ𝑣: Number of lockers with dimensions of h and v.
𝑈ℎ𝑣ℎ′𝑣′𝑠: Number of lockers (h,v) assigned to satisfy demand for locker of minimum dimension
(h’,v’) in scenario S.
Decisions to make (model’s decision variables)
Locker
configuration
Bank Dimensions
Interactive Module(s)
Location(s)
Capacity
Order
satisfaction

12. Optimization Based Design Model
12
Objective: Maximize profits generated by scenarios considered
max ෍
∀𝑠∈𝑺
𝑝𝑠 ෍
∀(ℎ,𝑣)∈𝑫 𝒔
𝑟ℎ𝑣𝑠 𝑆ℎ𝑣𝑠 − ෍
∀ ℎ,𝑣 ∈𝑫,∀ 𝑥,𝑦 ∈𝑮 𝒉𝒗
𝑐ℎ𝑣𝑥𝑦
𝐴
𝐴ℎ𝑣𝑥𝑦 − 𝑐 𝐻 𝐻 − 𝑐 𝑉 𝑉 − ෍
∀ ℎ,𝑣 ∈𝑫
𝑐ℎ𝑣
𝐿
𝑁ℎ𝑣
Revenues Generated Ergonomic Costs Manufacturing Costs

13. Optimization Based Design Model
13
Geometric constraints
Define the locker bank dimensions
(Grid used for the design)
Interactive modules constraint
Define the number of interactive modules
Insure a minimum distance 𝑚 between them
Locker configuration constraint
Define capacity and avoid superposition
Scenario’s order assignment constraints
Ensure unique order assignment

14. Optimization Based Design Model
14

15. Experimentation Method
15
“Standard” smart locker bank

16. Experimentation method
16
Normal distribution (μ,σ)
Small Medium Large
Delivery mix (S,M,L)
Optimization
model
Normal distribution (μ’,σ)
Small Medium Large
Delivery mix (S,M,L)’
Design scenarios
Experiment scenarios
Key Performance
Indicators1
2
3
4
5

17. Performance Analysis
17
400
450
500
550
600
650
700
750
800
1 10 19 28 37 46 55 64 73 82 91 100
Scenario
150
200
250
300
350
400
450
500
1 10 19 28 37 46 55 64 73 82 91 100
Scenario
Revenues
0
20
40
60
80
100
120
140
160
180
1 10 19 28 37 46 55 64 73 82 91 100
Scenario
Ergonomic costs Profits
(μ = 100,σ = 20) ; (S = 50%, M = 30%, L = 20%)

18. 0
10
20
30
40
50
60
70
80
90
100
ScenarioMean
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
105
110
115
120
125
130
20
30
40
50
60
70
80
90
100
ScenarioMean
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
105
110
115
120
125
130
0
100
200
300
400
500
600
700
800
ScenarioMean
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
105
110
115
120
125
130
Mean μ’ Sensitivity Analysis
18
Revenues Service Level Utilization
(μ = 100,σ = 20) ; (S = 50%, M = 30%, L = 20%)

19. Mean μ’ Sensitivity Analysis
19
Revenues Improvement Service Level Improvement Utilization Improvement
0%
10%
20%
30%
40%
50%
60%
25 35 45 55 65 75 85 95 105 115 125 135
0%
10%
20%
30%
40%
50%
60%
70%
25 35 45 55 65 75 85 95 105 115 125 135
0%
10%
20%
30%
40%
50%
60%
25 35 45 55 65 75 85 95 105 115 125 135
(μ = 100,σ = 20)

20. Mix (S,M,L)’ Sensitivity Analysis
20
Revenues Service Level Utilization
400
450
500
550
600
650
700
750
800
L = 60, M = 30,
S = 10
L = 50, M = 30,
S = 20
L = 33, M = 33,
S = 33
L = 20, M = 30,
S = 50
L = 10, M = 30,
S = 60
L = 5, M = 25,
S = 70
30
35
40
45
50
55
60
65
70
75
80
L = 60, M = 30,
S = 10
L = 50, M = 30,
S = 20
L = 33, M = 33,
S = 33
L = 20, M = 30,
S = 50
L = 10, M = 30,
S = 60
L = 5, M = 25, S
= 70
50
55
60
65
70
75
80
85
90
95
L = 60, M = 30,
S = 10
L = 50, M = 30,
S = 20
L = 33, M = 33,
S = 33
L = 20, M = 30,
S = 50
L = 10, M = 30,
S = 60
L = 5, M = 25, S
= 70
(μ = 100,σ = 20) ; (S = 50%, M = 30%, L = 20%)

21. Conclusions
21
• Length of stay of packages is not taken into account
• Package sizes are matching available lockers sizes

22. Questions
22
Thank you!
Louis.Faugere@gatech.edu