Mayurpankhi Barooah Seung Hwan Shin Agenda 1 Context Approach Analysis and Results Recommendations This project is aimed at optimizing multitemperature deliveries to smallformat stores Large format stores ID: 743663
Download Presentation The PPT/PDF document "Optimal Multi-Temperature Deliveries to ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Optimal Multi-Temperature Deliveries to Small-format Stores
Mayurpankhi Barooah, Seung Hwan ShinSlide2
Agenda
1
Context
Approach
Analysis and Results
RecommendationsSlide3
This project is aimed at optimizing multi-temperature deliveries to
small-format stores
Large format stores
Small format stores
A new DC to store delivery strategy is required for small format stores
1
Context
Average 180,000
sqft
High volume, frequent deliveries
Average 40,000
sqft
Higher proportion of grocery
Multi-temperature products
Smaller volume but require frequent delivery
Source: www.secardpools.com
Source:
www.myers-company.com
Source:
www.vanderbilt.edu
2Slide4
Multi-Temperature Trailers (MTT) could potentially help consolidate delivery volumes to stores
Key challenges in small format deliveries
Large number of stops
Limitations on allowable driving time
3
A cost benefit analysis was used to estimate advantages of using MTT
1
Context
Single Temp Trailer
Multi-
TempTrailer
Ambient
Frozen
Amb
Refrig
Frozen
Refrigerated
A
dvantages of Multi-Temp Trailers (MTT)
Consolidate demand onto one trailer
Fewer stops
Store 1
Store 2
Store 3
Store 1Slide5
We considered a sample of small format stores to identify the optimal delivery policy
4
Sample small format stores:
9 stores delivered by same
DC
Data collected:
Daily demand volume by product category
Distances Current delivery frequencies2
ApproachSlide6
We organized policy options by number of stops and single vs multi temperature trailers
Cost model helped identify optimal trailer
configuration, product mix and delivery
frequency
MSST
MSMT
SSST
SSMTSingleMultiSingle
MultiTemperatureStops
Policies
Minimize
total c
osts = Transportation costs+ Stoppage costs + Holding costs
Optimization
52
ApproachSlide7
C
ost, trailer utilization and delivery frequency were the key factors considered to evaluate delivery policies
6
Cost per
p
allet
:
Average per pallet cost (Transportation + Stop + Holding)
Trailer utilization rate:No. of pallets delivered per week / Total trailer capacity Delivery frequency:No. of deliveries to each store every week
Three Key Factors
3
Analysis and
Results
Key assumptions
Demand: Daily average in pallets for 3 months in
2014
Transportation and stop cost
Cost per mile for ambient and temp. controlledStoppage costInventory holding cost
Holding cost rateProduct value per pallet by categoryMinimum 4 deliveries a weekSlide8
For the ‘Base’ case, MSMT policy showed the lowest cost per pallet, the highest truck utilization rate and delivery frequency
7
Cost per Pallet:
$37
Utilization Rate:
37%
Delivery Frequency (
Wk):A 4.0 / F 4.0 / R 4.0Cost per Pallet: $17Utilization Rate: 85%Delivery Frequency (Wk):
A 5.2 / F 5.2 / R 5.2Cost per Pallet: $19Utilization Rate: 86%Delivery Frequency (Wk):A 4.1 / F 4.0 / R 4.1Cost per Pallet:
$16Utilization Rate: 94%
Delivery Frequency (Wk):
A 5.3 / F 5.2 / R 5.2
StopsTemperatureSingle
MultiSingle
Multi* A: Ambient / F: Frozen / R: Refrigerated 3
Analysis and ResultsSlide9
We tested 4 scenarios to
better understand which factors influence the policy
selection
8
Base Case
Doubled demand
1
Doubled distance
between DC and Stores2
7-day deliveries per Week
3
Half demand
4
3
Analysis and ResultsSlide10
The optimal policy changes from MSMT to MSST.
Higher demand drives the single-temp trailer more attractive.
For ‘Doubled demand’ scenario, the MSST becomes the most economical policy.
9
Doubled
d
emand
$13.96
92%A 4.6 / F 4.0 / R 4.3
$15.27
96%A 9.9 / F 9.9 / R 9.9
$19.3164%A 4.0 / F 4.0 / R 4.0$16.2490%A 10 / F 10 / R 10S
M
SMStopsTemperatureMSSTMSMT
SSSTSSMT3Analysis and Results
1
S
M
S
MStopsTemperature
MSST
MSMT
SSST
SSMTSlide11
For ‘Doubled distance’ scenario, the MSMT is the most economical policy.
10
Doubled
d
istance between DC and
s
tores
$30.5480%A 4.3 / F 4.0 / R 4.4$29.10
99%A 5.9 / F 5.9 / R 5.9
$68.51
37%A 4.0 / F 4.0 / R 4.0
$31.3385%A 5.2 / F 5.2 / R 5.2SM
S
MStopsTemperatureMSSTMSMTSSST
SSMTNo optimal policy changes from the base case.Longer distance from DC drives the multi-temp
trailer preferred as it increases utilization rates and minimizes the linehaul travel. 3
Analysis and
Results2
S
M
S
M
Stops
Temperature
MSST
MSMT
SSST
SSMTSlide12
For ‘7-day Deliveries’ scenario, the MSMT is the most economical policy.
11
Min. 7-day deliveries per week
$
27.21
60%
A
7.0
/ F 7.0
/ R 7.0
$
16.91
94%A 7.1 / F 7.1 / R 7.1$64.5820%A 7.0 / F 7.0
/ R 7.0
$22.4164%A 7.0 / F 7.0 / R 7.0SMSM
StopsTemperatureMSSTMSMTSSSTSSMT
No optimal policy changes from the base case.
The
higher delivery frequency target (higher freshness) makes the multi-temp trailer policy more attractive.
3
Analysis and Results
3
S
M
S
M
Stops
Temperature
MSST
MSMT
SSST
SSMTSlide13
For ‘Half Demand’ scenario, the MSMT is still the most economical policy.
12
Half
d
emand
$
30.86
53%
A 4.0 / F 4.0 / R 4.0$
18.20
93%A 4.7 / F 4.7 / R 4.7
$69.00
19%A 4.0 / F 4.0 / R 4.0$25.1957%A 4.0 / F 4.0 / R 4.0
S
MSMStopsTemperatureMSST
MSMTSSSTSSMT
No optimal policy changes from the base case.Smaller demand drives the multi-temp trailer more attractive. It needs to consider using smaller size trailers.3
Analysis and
Results4
S
M
S
M
Stops
Temperature
MSST
MSMT
SSST
SSMTSlide14
Summary - Cost per Pallet Comparison
13
3
Analysis and
Results
In
almost all scenarios, MSMT emerged as the lowest cost
policy.
The cost gap between single-temp trailer policies with multi-temp trailer policies was narrowed when demand increased.
($)Slide15
Demand, distance to stores and delivery frequency emerged as key determinants of delivery policy
Small demand: Multi-temp trailer
Large demand: Single-temp trailer
14
Consideration of
demand
Consideration of
distance between DC and stores
Longer distanceMulti-stops: Minimizing the linehaul tripsMulti-temp trailer: Increasing trailer utilization rate
Delivery frequency (Freshness)
Higher delivery
f
requency: Multi-temp trailer
4
RecommendationsSlide16
Future research can refine the existing model and address current limitations
15
Limitations
Variability
of daily
demand
Different holding cost rate
Inventory space constraints
Loading/unloading cost change
Future research
Incorporating limitations
Combinations of scenarios
(e.g. higher demand with longer distance between DC and stores)
Other
characteristics: Intra-zone distance, dramatically different volumes by store, different labor costs, etc.
‘Flex-temp’4RecommendationsSlide17
16
Questions?Slide18
Back up
17
Assumptions
Summary of results
Optimization modelSlide19
Assumptions
18
Parameter
Definition
Value
Unit
Trailer cost per mile
Cost incurred by a trailer depending on trailer typeAmbient: 2.35Temperature controller trailer (Frozen or refrigerated): 2.72$ per mileStop cost
Cost paid to the carrier for every stop50$ per stopProduct valueAverage product value assumed for inventory cost calculationAmbient: 560Frozen: 720Refrigerated: 640$ per palletHolding cost
Annual holding cost rate
25
%
Trailer TypeMaximum WeightMaximum Cubic FeetMaximum No. of Pallets
Ambient temperature trailer
47,0002,10028Refrigerated and Frozen temperature trailer41,0001,80028Multiple temperature trailer41,0001,80026Slide20
Summary of results (1/3)
19
Scenario
SSST
SSMT
MSST
MSMT
Base37171916
Doubled demand19161415Doubled distance693131
29
7 day delivery
6522
2717Half demand6925
31
18Cost per Pallet by Policy ($)ScenarioSSSTSSMTMSSTMSMTBase231%106%
119%100%Doubled demand127%114%100%107%Doubled distance238%107%107%
100%7 day delivery382%129%159%100%
Half demand
383%139%172%100%Cost per Pallet Comparison to Optimal PolicySlide21
Summary of results (2/3)
20
Utilization by policy (%)
Scenario
SSST
SSMT
MSST
MSMTBase378580
93Doubled demand64909296Doubled distance3785
80
997 day delivery
20
646094Half demand19
5753
93ScenarioSSSTSSMTMSSTMSMTBase40%91%
86%100%Doubled demand67%94%96%100%Doubled distance37%86%81%
100%7 day delivery21%68%64%100%
Half demand
20%61%57%100%Utilization Comparison to Optimal PolicySlide22
Summary of results (3/3)
21
Scenario
Product
SSST
SSMT
MSST
MSMTBaseAmbient45.2
4.15.3Frozen 45.245.2Refrigerated
45.2
4.1
5.2
Doubled demandAmbient410.14.6
9.9
Frozen 410.149.9Refrigerated410.14.39.9Doubled distance
Ambient45.24.35.9Frozen 45.245.9Refrigerated
45.24.45.97 day delivery
Ambient7
777.1Frozen 7
77
7.1Refrigerated777
7.1
Half demand
Ambient
4
4
4
4.7
Frozen
4
4
4
4.7
Refrigerated
4
4
4
4.7
Delivery Frequency (Number of Deliveries per Week) by PolicySlide23
Model:
Objective
22
Optimization Model Developed
by Unahalekhaka (2015)Slide24
Model:
Constraints (1/4)
23
Optimization Model Developed
by Unahalekhaka (2015)Slide25
Model: Constraints
(2/4)
24
Optimization Model Developed
by Unahalekhaka (2015)Slide26
Model: Constraints
(3/4)
25
Optimization Model Developed
by Unahalekhaka (2015)Slide27
Model: Constraints
(4/4)
26
Optimization Model Developed
by Unahalekhaka (2015)Slide28
Model Summary (1/3)
27Slide29
Model Summary
(2/3
)
28Slide30
Model Summary
(3/3
)
29