Distribution network Flashcards
1
Q
Logistic network is the combination of
A
Supply network
Distribution network
2
Q
From Suppliers to Manufacturers
Tipically convergent
A
Supply network
3
Q
From Manufacturers to Customers
tipically divergent
A
Distribution network
4
Q
Design principles can be apply same way for Supply than for Distribution networks
A
yes
5
Q
What are the design principles for a DN
A
- problem setting
- decision framework
- distribution problem profile
- typologies of distribution networks
- Design principles
6
Q
Decisions Variables od a DN
A
number of echelons
type
number of nodes
location
7
Q
Contextal variables f a DN
A
Sourcing Strategy
Sales channel
Product Feature
Objective function
8
Q
How do I select the best alternative out of the many posibilities of a DN
Best DN at target SL & min Cost
A
You filter by strategic design (Qualitative cost models) selecting the most promising to apply quantitative models
9
Q
What is a distribution system
A
all resources and structures involved in transportation of goods
10
Q
Main decision variables
Strategic vs operational
Network:
number of echelons, number of WH & location, automation level, MVB
Transport:
Mode
MVB
Inventories:
Inventory management
SS allocation
Centralized/Decentralized
A
Strategic Level
11
Q
Main decision variables
Strategic vs operational
Network:
product allocation, storage hanling systems, order making, connections
Transport:
Delivery lots sdimensioning
local distribution management
Inventories
Order making
time btwn orders
demand forecasting
back order management
A
Operational
12
Q
Product range
Density
Value Density
Obsolescence
A
Product features
13
Q
min(logistic cost), Service Level constrain
max(service level), logistic cost constrain
min(logistics costs + poor-service cost)
A
Objective function
14
Q
poor service cost
A
Stock out
Bad quality
missed delivery
15
Q
Distribution cost
A
- Transportation cost
Primary
Secondary - Inventory
Cycle stock
Safty Stock
In-transit stock - Handling cost
- Other
16
Q
Distribution Cost
Transport
from suppliers to the central warehouses and from the central warehouses to the regional warehouses/transit points
A
Primary Transport
17
Q
Distribution Cost
Transport
Local distribution (usually from the distribution network nodes to the end customers)
A
Secondary transport
18
Q
Distribution cost
Inventory
these inventories deal with the different operative rhythm of two following stages in the supply chain
A
Cycle stock
19
Q
Distribution cost
Inventory
these inventories deal with the uncertainty of both the demand and the replenishment lead times
A
Safty Stock
20
Q
Distribution cost
Inventories
these inventories are in transit between stocking or production points (mainly inside the vehicles)
A
In transit
21
Q
Distribution cost
due to the handling activities in the warehouses and the transit points of the distribution network (loading/unloading, picking, etc.)
A
Handling
22
Q
Distribution cost
order management, packaging, reverse logistics
A
Other
23
Q
Distribution costs
What the cost categories should consider?
A
How these costs vary with the main features of the distribution network.
24
Q
Distribution cost main driver
Primary transportation
A
Saturation
Distance
Flow related to shipment
25
# Distribution cost main driver
Secondary transportation
Points Saturation
Tour avg distance
26
# Distribution cost main driver
Safty Stock
Demand and lead time uncertainties, average lead time
27
# Distribution cost main driver
Cycle Stocks
Re-order frequency between nodes along the distribution network
28
# Distribution cost main driver
In transit stock
Flow and total in transit
29
# Distribution cost main driver
Handling
Flow
30
# Customer service
Time elapsed between the order issue and the delivery of goods
Order Cycle Time OCT
31
# Customer service
To deliver the order in the time window (ΔT) that has been arranged with the customer
Delivery puncltuality
32
# Customer service
All the goods included in the order are delivered in a single delivery
Delivery completness
33
# Customer service
To satisfy urgent/unexpected needs of the customer
Delivery flexibility
34
# Customer service
35
# Customer service
Number of deliveries scheduled in the reference time window (day, week,…)
Delivery frequency
36
Potential roles of central warehouses in a 1-
echelon network
reduction of the order cycle time
product mixing
optimization of transport (N+M)
Centralize SS
37
Distribution problem to tacle with a 1 echelon
POO far from markets
OCT stringent
Specialized POO & many
Completeness important KPI
Many connections (POO POD)
High incidence in transport cost
38
Role of 1 level network
↓ Order Cycle Time
↑ Order Completeness
↓ Transportation cost
39
How do we decreade the incidence in transportation cost with 1 level network?
By decreasing amount of connections, saturation of the trucks and slighly increase of distance
40
# Design Principles
____ should be considered as a baseline solution when no one of the 3 potential functions of L1 are very important
Direct delivery
41
# Design principles
____ of the network has 3 main potential functions – reducing the OCT, increasing order completeness, reducing (primary) transportation costs. If at least one of these 3 functions is very important, then level 1 should be considered as the baseline solution.
Level 1
42
# Width of L1/ Number of nodes at L1
The number of warehouses at L1 depends on:
* OCT: the shorter the more we need
* Transportation and inventory cost trade off
43
# Design principles
The ____ depends on a trade-off between inventory-related costs and transportation costs subject to service level constraints
number of nodes at L1
44
What are the drivers behind a different number of WHS at the first level of the network?
Logistic cost:
Inventory
Transport cost
OCT
Service level
45
Descrive the Logistics cost graph for a level 1
Logistics cost is conformed by Inventory related and transportation relted. Inventory rise when WH increase, and primary transportation cost decreases when WH increases.
The optimarl N of WH is the intersection of this two curves setting a minimum in the LC.
46
I will need few WH when:
High value density
High risk of obsolescence
High product range
47
I will need many WH when
Low value density
low risk of obsolescence
low product range
48
Descrive the OCT graph for a L1
The OCT decreases when the SL increases and so number of WH
49
Even though the intersection of inventory qnd hqndling gives the optimal value, the SL for that WH number is not the one we need...
What should we do?
We should define the Min OCT to relate it to the number of WH.
So that we get the Number that minimaze and gives us the SL we need.
50
They aim to optimize transports to the end customers (local distributio) by accepting a longer cycle time (+ 1 or 2 days on average)
operative models:
only sorting of Transport Units which have been completely picked upstream
receipt of a massive delivery (= aggregation of different customer orders), picking and consolidation of Transport Units for each customer
Transit Points
51
Why would I choose a WH instead of a TP in a L1
Because decreasing the OCT is the main important thing
52
# Design principle
Nodes at L1 should be WHS if reducing the OCT is a very important function of L1, otherwise consider TPs.
| Just a reminder
Types of nodes
53
The **second tier of warehouses** in a 2-echelon distribution network **provides the following supplementary functions**:
1. **Higher service level **in terms of both cycle time and punctuality
(thanks to the inventory in the regional warehouses)
2. **Optimization of transport** to the end customers (local distribution)
54
Role of level 2 (L2 vs L1) What is the "role" (value, potential) of the second level in a distribution network?
↓ ↓ Order Cycle Time
↓ Secondary Transport cost (from L1 customers)
55
Distribution problem Profile for a L2
We need veeery short deliveries, short OCT day-hrs
We have many points of delivery ordering small quantities frequently
56
The number of nodes at L2 depends on the trade-off between
Inventory-related costs and secondary transport cost, subject to service-level constraints.
In this case we will take the # WH that place us in a range to reach the pints in <1hr this is hhow we consider the service level in this example
57
If both functions of L2 are very important, than consider L2 with WHS.
Otherwise, if the important function is to reduce ... consider TP
Secondary Transportation cost
58
What can we do if we want to increase network flexibility
Implement a mixed network deliveries to all levels possible
59
... are a solution to distribution problems having more than one distribution problem profile at the same time
Mixed networks
60
# true or false
Level 3 of the distribution network should be considered only when there are very challenging requirements in terms of order time reduction and/or transportation costs optimisation
True
61
Potential Value of outsourcing Logistics
Less investment
More focus in core activities
Increases Economies of scale
Increases Innovation knoledge
increases Flexibility
62
Potential Risk and Cost of outsourcing Logistics
Less control
Less competence
Coordinating cost
Switching cost
63
How can I get from L0 to best alternative of a Distribution Network
By strategic Design first filter "Qualitative Cost Models" then from those options I apply Mathematical models and select Best alternative
Best solution at needed SL and min Cost
64
What do we mean by strategic design filter when definin DN
Qualitative Models Design Principles
Qualitative Cost Models
65
What are the mathematical models to choose a DN
Quantitative Cost Model
Center of Gravity
Linear programming
Mixed Integer Proframming
66
They aim at analyzing the relationships between the main costs and a specific parameter of the Distribution Network (e.g. number of warehouses, number of echelons, etc.) by identifying the cost curves (i.e. curves that show how the costs vary by varying the parameter).
Qualitative Cost Models
67
They aim at calculating the costs of one (or a limited number of) specific Distribution Network configuration in order to assess the most convenient alternative. They require a huge amount of data and can be used only after the most promising alternative selection.
Quantitative Cost Models
68
Is a function of **distance** between Central WH and Regional WH and the saturation (flow/shipment size)
Primary Transportation
69
Full transport truck load will hold steady until which point?
Untill the number of regional WH impact in a decrease of saturation of the Trucks.
Increasing: fare, frequencies
70
What happens if the number of regional warehouses increases?
Distance RWH area will decrease
Delivery area inncreases by truck
This process will keep until the distances between regional warehouses and delivery areas will be irrelevant compared to the distances traveled in the delivery area (dA).
= having a local distibutor
71
If the replenishment frequency does not vary (i.e. time lapse between two consecutive orders does not change), the ____ will not change by varying the number of regional warehouses.
(If the replenishment frequency changes, the overall ____ in the system will vary – it is proportional to the **time between consecutive orders** –)
Cycle Stock
CS=0.5xDemandxT
T: Timelapse between two consecutive Orders
72
If RWH increases Cycle Stocks:
Remains the same they depend on demand and time between two consecutive orders
73
If Demand changes, CS:
Changes as well
74
The overall amount of the ____ only depends on the **replenishment lead time and the demand.** Neither the lead time nor the demand depend on the number of warehouses.
In transit stocks
ITS= LT x D
| LT: time from order to inboud
75
____ in the system depend on different factors: * service level
* demand seen by each warehouse and its standard deviation
* geographical correlation of the demand
* replenishment LT and its standard deviation
* inventory policy
____ deal with the **unpredictability** of both **demand and LT**. The more the demand is split among warehouses, the more the ____ in the system.
Safty Stocks
Reorder point model:
K factor times the Square root of the sum of the product between ELTxD2sd and ED2xLT2sd
Periodic rivew model:
ELT + T....
76
Due to economies of scales which inventories coul be affected (slighly increasing from range to range of RWH)
In terms of cost to store one unit
Cycle stock
Safty Stock
77
The ____ (the cost of loading/unloading the vehicles, handling, picking, etc.) is subject to economies of scale (for instance automated warehouses only for high flows).
The handling cost per unit
78
How Handling cost is affected if same technology among WH and how if different
Remains constant
Increases by tech
79
Location
To find the exact position
Site selection
80
Allocation
designate the flow to be allocated to each warehouse (i.e. the quantities on the “arcs” of the network)
the customer allocation to the warehouses
81
This process is based on
* the determination of the geographical position of the warehouse
* the review of the results stemming from the quantitative techniques by taking into account more specific factors
the location process
82
Quantitative techniques fot WH location
1. Single WH
Center of Gravity
2. Multi-WH
Heuristic models
Linear programming (simple - mixedi nteger)
Simulation
83
Choice factors for the location problem
* **Proximity** to suppliers/plants
* **Proximity** to customers
* Presence of transport **infrastructures**
* **Cost** of the **area** and of the public utilities
* **Costs** of **in/outbound** transports
* Cost and r**eliability of labor**
* ** Ta x** r e l i e v e s / r e s t r i c t i o n s
* **Proximity** to other companies
* Meteorological **conditions** / quality of **life**
84
Single WH location, isue and relevant cost:
The only issue is where to locate one warehouse. The most relevant costs are
transportation costs
85
Multi WH location, problem and issues:
It’s a problem of both **location** and **allocation** (location-allocation problem).
A multi-warehouse location model has to deal with the following issues:
Multi-warehouse **location**
**Product and production capacity allocation** to both plants and warehouses
**Customer allocation** to the warehouses
86
Explain the Model classification matrix
(Location,Transportation Cost)=Center of gravity
(Allocation, Transportation+Handling+Inventory cost)= Linear programming
Location+allocation, Transportation+Handling+Inventory+...+Customer service)= Mixed - integer programming
87
Center of Gravity:
Application field
Main Managerial issue
Model Parameters
**Application field**
location of a single warehouse
**Main managerial issue**
(X,Y) = coordinates of the center of gravity
**Model parameters**
(X i,Yi) = coordinates of both the points of origin and destination
Fi = inbound (for the point of destination) and outbound (for the point of origin) flows
Ri = transportation rate per unit [€/(km*t)] (it depends on the weight and the distance)
88
Center of Gravity
Objective function
Minimization of the transportation costs by finding the coordinates of the center of gravity.
Transportation costs linearly depend on the distance and the transported quantity
min( Total Cost) = minSum[FxRxd(x,y)]
| F: Flow
R: Rate
d:distance
sum of all POO and POD trated the same way
89
First step of Center of Gravity's process
I will randomly choose a first rough (x,y) and calculate:
x* = SFi x Sxi / SFi
y* = SFi x Syi / SFi
I get (x*, y*) the centroid
| all in sums!!!
Normally I should include the Rate but is ok
90
Once you calculated the first Centroid what can we do:
As distances can be calculated we have: Flows, distances and by ende fares. We calculate TC and then we proceed for a second iteration.
91
How is the second iteration after the calculation of the first centroid done:
We can re calculate the position by:
x* * = (S(Fi x Ri x xi) / di) / (S(Fi x Ri) / di)
y* * = (S(Fi x Ri x yi) / di) / (S(Fi x Ri) / di)
By now we. recalculate TC= S(FixRiXdi)
we iterate until the TC doesnt change much
92
How can we find the distances di for the center of Gravity
By Euclidean distances (sqrroot of the sum of squres od difference of x and y). Outside cities
Reclilinear distances (sum of absolutes of diferences of x,y). Cities
93
Center of gravity key messages
* The result does not depend on the origin (0,0)
* In the first step, it is strongly recommended to use a constant Ri (e.g. Ri=1 ∀i)
* It is a very useful method to find a rough location (even though it doesn’t include some costs e.g. handling costs, storage costs, etc.)
* It can be used also to find isocost curves in order to assess the trade off with the other costs
94
Linear programming
Application field
Model Variables
Models parameters
**Application field **
Allocation of the plant flows to the warehouses and/or allocation of the customers to the warehouses
**Model variables **
xij = quantity to be delivered from the plant i to the warehouse j
yjk = quantity to be delivered from the warehouse j to the customer k
**Model parameters **
Rij = Transportation fare per unit from the plant i to the warehouse j [€/unit]
Rjk = Transportation fare per unit from the warehouse j to the customer k [€/unit]
Qi = **production capacity** of the plant i
Qj = **operative capacity** of the warehouse j
Qk = **demanded quantity** by the customer k
95
Linear programming Objective function and constrains
Minimize the transportation costs by determining the **optimal flows** from the plants to the warehouses and from the warehouses to the customers
Constraints less or equal than real!
* production capacity of the plants (Qi)
* operative capacity of the warehouses (Qj)
* demand of each customer (Qk)
96
How does the clustering method works
by alocating one wh for customer assesing cost. reduce wh by one and apply center of gravity. continue until the cost does not increase or you have just one WH.
97
MIP Mixed Integer Programming
aims to reduce:
* production cost
* fixed WH cost
* Variable WH cost
* Primary and Secondary transportation
The objective is to minimize total costs, which is the sum of:
Fixed costs for opening warehouses at certain locations.
Variable costs for transporting products from warehouses to customers.
So, we want a solution that keeps both these costs as low as possible.
98
Constrains of MIP
Capacity Constraints: Each warehouse can only hold a certain amount of product.
Demand Constraints: Each customer needs a specific quantity of product.
Warehouse Status Constraints: Products can only flow from a warehouse if it's open.
99
How MIP Solves It
MIP algorithms (like branch-and-bound) try various combinations of warehouse locations and allocations, keeping track of those that meet all constraints and cost less.
The best (lowest-cost) solution that meets all requirements is the optimal solution, where we know where to open warehouses and how to allocate products efficiently.
100
MIP variables
Binary variables that indicate if a warehouse is open at a certain location (1 = open, 0 = closed).
Integer variables that show how much product flows from each warehouse to each customer (usually non-negative integers).
