Over. Def. App.
Distribution and Network Models
Master of Business Administration
MBA 8419 – Decision Making Technology
Copyright By PowCoder代写 加微信 powcoder
1 Distribution and Network Models
Over. Def. App. Overview of the presentation
Definitions Network
General Optimization Model
Applications
Multi-period planning
General principles
Production planning basic case
Production planning with general deliveries
Logistics and transportation Transportation problem
Distribution and Network Models
Over. Def. App. Net. Flo. Gen. Opt. Mod.
Definitions Network
Defined using a graph, which is a structure defined as a set of nodes for which some pairs of the nodes are connected via arcs.
Arc (i, j) : where i = initial node and j = terminal node
(i,j) ⇒ emerges (leaves) node i and is incident to (arrives) at node j
Arc : defines a specific relationship between two nodes Examples
route linking intersection i to j assignment of employee i to task j renting a vehicle i to a client j
Distribution and Network Models
Over. Def. App.
Net. Flo. Gen. Opt. Mod.
Definitions Network
FIGURE – Visualization of social network analysis
4 Distribution and Network Models
Over. Def. App. Net. Flo. Gen. Opt. Mod.
Definitions Flow
In a network, flow refers to the units (e.g., goods, materials, people, etc.) that move on the arcs following their specific direc- tion.
Associated with each arc (i,j)
xij = the number of units (i.e., quantity of flow) that move
along the arc (i, j)
cij = unit cost for the flow moving along (i,j) Bounds on the quantity of flow associated with xij
uij = maximum quantity of flow that can move along arc (i,j) lij = minimum quantity of flow that can move along arc (i,j)
lij ≤ xij ≤ uij
Distribution and Network Models
Over. Def. App. Net. Flo. Gen. Opt. Mod.
Definitions Flow
Flow (cont’d) :
Associated with each node i
b(i) = demand value associated with i
There are three possible cases with respect to the demand values :
b(i) > 0 ⇒ node i defines an origin for the flow (i.e., flow enters the network at this node).
b(i) < 0 ⇒ node i defines a destination for the flow (i.e., flow leaves the network at this node).
b(i) = 0 ⇒ node i is a transhipment node (i.e., flow simply transits at this node and remains within the network).
If the values b(i), for all nodes i, are integer, then solution to the network flow problem will also be integer (i.e., without the need to impose the integrality requirements).
6 Distribution and Network Models
Over. Def. App. Net. Flo. Gen. Opt. Mod.
Definitions
General Optimization Model
Decision variables
xij = number of units of flow that transit on arc (i,j) Objective Function
Total cost incureed to distribute the flow through the network
cijxij for all arcs in the network (i,j)
Subject to
Flow conservation constraints at all nodes
For each node i → total flow on the arcs leaving i - total flow on the arcs arriving at i = b(i)
Bounds on the flow transiting through each arc Foreacharc(i,j)→lij ≤xij ≤uij
Distribution and Network Models
Over. Def. App.
Net. Flo. Gen. Opt. Mod.
Definitions
General Optimization Model
General Example :
(a) Costs, bounds, b(i) (b) Network FIGURE – Network flow problem
8 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Multi-period planning
General Principles Single period planning
Consists of decisional problems that occur for a single moment in a time horizon and that only considers the ressources available (supply) and state of the market (demand) for that particular mo- ment.
Multi-period planning
Consists of decisional problems that occur over multiple moments in a time horizon and that explicitly take into account the dynamic by which available ressources and market conditions can evolve (i.e., change) through time.
9 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Multi-period planning
General Principles (cont’d)
FIGURE – Single period planning process
10 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Multi-period planning
General Principles (cont’d)
FIGURE – Multiple periods planning process
11 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Multi-period planning
Production planning basic case : Pastissimo inc.
Pastissimo is an Italian company that specializes in the production of high- quality pasta for a variety of clients. The company has recently received an important order from one of its client, Hyper-Halli. Following this order, for the next 6 months, Pastissimo will deliver (in 1kg bag units) the spaghetti that is sold by Hyper-Halli as its own brand. Therefore, at the end of each month, Pas- tissimo will deliver 4 tons of spaghetti to Hyper-Halli, which has agreed to pay 5.28$ per bag for these deliveries. The production of spaghetti requires the use of wheat. To ensure that enough wheat will be available, Pastissimo has nego- tiated a contract with a local producer. The details of the contract are provided in the following table :
Month Price in $ tà Minimum in tà Maximum in tà ààà 4 6
2 9à5 3 4
3 ààà 5 à
4 98à 2 3 5 à2à 4 à 6 à25 5 6
FIGURE – Contract with wheat producer
12 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Multi-period planning
Production planning basic case : Pastissimo inc. (cont’d)
To store either the wheat that is bought from the producer or the spaghetti that is produced, Pastissimo has both a silo, where wheat can be stockpiled, and a warehouse, where the final products can be kept. At the beginning of month 1, the silo already has 2 tons of wheat and the company would like to keep the same amount at the end of the 6th month. The silo can store up to 3 tons of wheat and the monthly storing cost is 20$/t. As for the store, its capacity is 1 ton of spaghetti and the monthly storing cost is 25$/t. To ensure that Pastissimo delivers the required amounts of spaghetti to Hyper-Halli for the next 6 months, the manager planned the production capacity and costs as follows :
Month Production Capacity in tà Production Costs in $ tà 6 6à
4 4 6à 5 4 à5 6 3 65
FIGURE – Production capacity and costs
13 Distribution and Network Models
Over. Def. App. M.-P. Pla. Log. Tra.
Applications Multi-period planning
Production planning basic case : Pastissimo inc. (cont’d) Question
Pastissimo is interested in planning its operations to perform the order to Hyper-Halli for the next 6 months.
Therefore,
The company is looking to determine the following :
Supply of wheat
Inventory (wheat and spaghetti) Production levels
Distribution and Network Models
Over. Def. App. M.-P. Pla. Log. Tra.
Applications Multi-period planning
Production planning basic case : Pastissimo inc. (cont’d) Network flow model
Identify the Beginning and End of each month
Bi, i = 1,...,6 Ei, i = 1,...,6
Supply of wheat : • → Bi Production : Bi → Ei
Storing wheat : Bi−1 → Bi Storing spaghettis : Ei−1 → Ei
Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Multi-period planning
Arcs i, j Costs cij Bound lij Bound uij •,B ààà 4 6 •,B2 9à5 3 4 •,B3 ààà 5 à •,B4 98à 2 3 •,B5 à2à 4 à •,B6 à25 5 6 B,E 6à à 6 B2,E2 5à à 5 B3,E3 5à à 4 B4,E4 6à à 4 B5,E5 à5 à 4 B6,E6 65 à 3 B,B2 2à à 3 B2,B3 2à à 3 B3,B4 2à à 3 B4,B5 2à à 3 B5,B6 2à à 3 E,E2 25 à E2,E3 25 à E3,E4 25 à E4,E5 25 à E5,E6 25 à
b B b B2 b B3 b B4 b B5 b B6 2 à à à à -2 b E b E2 b E3 b E4 b E5 b E6 -4 -4 -4 -4 -4 -4
(a) Costs, bounds, b(i)
FIGURE – Network flow for Pastissimo
(b) Network
16 Distribution and Network Models
Applications Multi-period planning
Over. Def. App. M.-P. Pla. Log. Tra.
Decision variables
X•Bi = number of t. of wheat that are bought at the beginning of month i, where i = 1,2,...,6
XBi Bi +1 = number of t. of wheat that is stored in the silo from the beginning of month i to the beginning of month i + 1, where
i = 1,2,...,5
XBi Ei = number of t. of spaghetti produced during the month i , where i = 1,2,...,6
X +1 = number of t. of spaghetti that are stored in the warehouse from the end of month i to the end of month i + 1, where i = 1, 2, . . . , 5
Distribution and Network Models
Over. Def. App. M.-P. Pla. Log. Tra.
Applications Multi-period planning
Objective function
min1000X•B1 +975X•B2 +...+20(XB1B2 +XB2B3 +...+XB5B6)+ 160XB1E1 +150XB2E2 +...+25(XE1E2 +XE2E3 +...+XE5E6)
Subject to
Flow conservation constraints at all nodes For example
NodeB4 →XB4B5 +XB4E4 −X•B4 −XB3B4 =0 NodeE2 →XE2E3 −XB2E2 −XE1E2 =−4
Bounds on the arcs For example
4≤X•D1 ≤6 0≤XD5F5 ≤4 0≤XF2F3 ≤1 etc.
Distribution and Network Models
Over. Def. App. M.-P. Pla. Log. Tra.
Applications Multi-period planning
Production planning with general deliveries
Assuming the Pastissimo renegotiates its distribution contract with Hyper-Halli, which now accepts advanced deliveries, or, late deliveries. Specifically, the new contract allows the following delivery options :
Late deliveries by one month can be accepted by Hyper-Halli, provided that Pastissimo pays a fee of 35$/t. for all spaghetti that is delivered late.
Advanced deliveries by one or two moths can be accepted by Hyper-Halli, provided that Pastissimo pays a fee of either 14$/t. or 17$/t. for all spaghetti that is delivered in advance by one and two months, respectively.
In addition, each time a bag of spaghetti is delivered to Hyper-Halli, Pastissimo pays a cost of 0.05$/kg. in transportation fees.
Question : How can the previous network flow model be adjusted to represent this new situation ?
19 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Multi-period planning
Arcs i, j Costs cij Bound lij
E,D 5à à E,D2 64 à E,D3 6à à E2,D 85 à E2,D2 5à à E2,D3 64 à E2,D4 6à à E3,D2 85 à E3,D3 5à à E3,D4 64 à E3,D5 6à à E4,D3 85 à E4,D4 5à à E4,D5 64 à E4,D6 6à à E5,D4 85 à E5,D5 5à à E5,D6 64 à E6,D5 85 à E6,D6 5à à
Bound uij à
b B b B2 b B3 b B4
2 à à à à -2
b E b E2 b E3 b E4 b E5 b E6 àààààà b D b D2 b D3 b D4 b D5 b D6 -4 -4 -4 -4 -4 -4
(a) Costs, bounds, b(i)
FIGURE – Network flow for Pastissimo with generalized deliveries
(b) Network
20 Distribution and Network Models
Over. Def. App. M.-P. Pla. Log. Tra.
Applications Logistics and transportation
General Context
Transportation is at the heart of logistics operations and one of the major drivers of economic activities. People and goods need to be efficiently moved throughout the world in order for societies and economies to function and thrive.
Transportation Problem
Base case :
Adopting the point of view of either the shipper or the receiver
Specific detailed routes are not considered
Service from origin-destination and the overall cost is important
Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Logistics and transportation
Transportation Problem (cont’d)
Consider the problem of a company who needs to supply its ware- houses with finished products that are then distributed to clients. The products are produced at a series of plants and, at the end of each month, they are transported towards the different ware- houses of the company.
For the next month, the company needs to perform the following operations :
Chicago Kansas City Houston 120 u. 80 u. 80 u.
Warehouses
Atlanta Los Angeles 150 u. 60 u. 70 u.
22 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Logistics and transportation
Transportation Problem (cont’d)
The cost of shipping between cities is usually a function of, given a set of possible services, min distances between the cities × a tarif per unit.
Assuming that the following unit costs (i.e., $/unit) apply :
Chicago Kansas City Houston
Atlanta Los Angeles 865
15 12 10 3 10 9
Question : How can this problem be formulated as a network flow problem ?
23 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Logistics and transportation
FIGURE – Illustration of the transportation problem
24 Distribution and Network Models
Applications Logistics and transportation
Over. Def. App. M.-P. Pla. Log. Tra.
Defining the network
2 types of nodes (Plants and Warehouses) :
Pj,wherej =1→Chicago,j =2→KansasCity,j =1→ Houston
Wi,wherei =1→NewYork,i =2→Atlanta,i =3→Los Angeles
Arcs represent the transportation of units Plants → Warehouses
Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Logistics and transportation
(a) Transportation problem network
FIGURE – Network flow - Transportation problem
minz = 8xPW + 6xPW2 + 5xPW3 + 5xP2W + 2xP2W2 + àxP2W3 + 3xP3W + àxP3W2 + 9xP3W3
xPW + xPW2 + xPW3 = 2à xP2W + xP2W2 + xP2W3 = 8à xP3W + xP3W2 + xP3W3 = 8à
xPW xP2W xP3W = 5à xPW2 xP2W2 xP3W2 = 6à xPW3 xP2W3 xP3W3 = àà
(b) Optimization model
26 Distribution and Network Models
Over. Def. App. M.-P. Pla. Log. Tra.
Applications Logistics and transportation
Assignment problem
The assignment problem is a special case of the transportation problem.
Definition
The problem instance has a number of agents and a number of tasks. Any agent can be assigned to perform any task, incurring some cost that may vary depending on the agent-task assignment. It is required to perform all tasks by assigning exactly one agent to each task and exactly one task to each agent in such a way that the total cost of the assignment is minimized.
27 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Logistics and transportation
Assignment problem (cont’d)
Assuming that in the previous context, the company was looking to assign a single production plant to a single warehouse to perform the necessary supply activities by simply considering the unit transportation costs.
minz = 8xPW + 6xPW2 + 5xPW3 + 5xP2W + 2xP2W2 + àxP2W3 + 3xP3W + àxP3W2 + 9xP3W3
xPW + xPW2 + xPW3 = xP2W + xP2W2 + xP2W3 = xP3W + xP3W2 + xP3W3 =
xP W xP 2W xP 3W = xP W 2 xP 2W 2 xP 3W 2 = xP W 3 xP 2W 3 xP 3W 3 =
(a) Assignment problem
FIGURE – Network flow - Assignment problem
(b) Optimization model
28 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Logistics and transportation
Assignment problem (cont’d)
Intercity truck transportation
Distances (km) :
Loads 1234567
1 Scranton 2 Honesdale 3 Y NY Dover Paterson 229 229 139 176 212 212 114 155 111 111 32 54
Newton 146 116 125 153 123 91 108 81 25
4Edison 62 62 69 68
5 Princeton 6 Warwick 7 Newark
Question :
92 92 84 95 116 116 62 69 54 54 43 26
88 89 111 44 101 76
How should the company proceed to solve this transportation problem ?
29 Distribution and Network Models
Over. Def. App.
M.-P. Pla. Log. Tra.
Applications Logistics and transportation
FIGURE – Network - Intercity truck transportation problem
30 Distribution and Network Models
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com