Optimized deployment of parts in a distribution network

0Associated
Cases 
0Associated
Defendants 
0Accused
Products 
1Forward
Citation 
0
Petitions 
10
Assignments
First Claim
1. A computerimplemented system deploying parts in a distribution network, the system comprising:
 a computer coupled with one or more databases, the computer configured to;
access data associated with a plurality of parts at a plurality of locations in the distribution network;
calculate a demand for each part at each location;
estimate an availability leadtime for each part at each location; and
generate a coverage function for each location based on the availability leadtime and the accessed data for each part.
10 Assignments
0 Petitions
Accused Products
Abstract
A method for deploying parts is disclosed. Locations that include supply locations and demand locations are defined. A supply location supplies parts to a demand location. A demand is computed for each part at each location. An availability leadtime is estimated for each part at each location. A leadtime demand is computed for each part at each location using the availability leadtimes for the part. A stock level is computed for each part at each location. A completely filled demand is determined from the leadtime demands and the stock levels, and a partially filled demand is determined from the leadtime demands and the stock levels. A coverage function for the parts at the locations is generated from the completely filled demand and the partially filled demand.
36 Citations
View as Search Results
Iterative measures  
Patent #
US 9,195,690 B2
Filed 06/26/2013

Current Assignee
SAP SE

Sponsoring Entity
SAP SE

Optimized deployment of parts in a distribution network  
Patent #
US 7,672,867 B2
Filed 10/22/2007

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
JDA Technologies US Inc.

Method and apparatus for scheduling work orders in a manufacturing process  
Patent #
US 7,519,444 B2
Filed 08/08/2006

Current Assignee
Infor Amersfoort BV

Sponsoring Entity
Infor Global Solutions BV

Optimized deployment of parts in a distribution network  
Patent #
US 7,337,031 B1
Filed 10/25/2001

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
i2 Technologies US Incorporated

Multiplyintegrated system for product inventory, sales, and distribution  
Patent #
US 7,363,249 B1
Filed 06/04/2001

Current Assignee
MOONSHADOW TECHNOLOGY CORPORATION

Sponsoring Entity
MOONSHADOW TECHNOLOGY CORPORATION

Redistribution of parts in a distribution network  
Patent #
US 7,210,624 B1
Filed 10/25/2001

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
i2 Technologies US Incorporated

Orderbased material management system  
Patent #
US 6,516,301 B1
Filed 05/03/1999

Current Assignee
WSOU Investments LLC

Sponsoring Entity
AlcatelLucent USA Inc.

System and method for extended enterprise planning across a supply chain  
Patent #
US 5,974,395 A
Filed 08/21/1996

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
i2 Technologies Incorporated

Optimization of manufacturing resource planning  
Patent #
US 5,630,070 A
Filed 08/16/1993

Current Assignee
International Business Machines Corporation

Sponsoring Entity
International Business Machines Corporation

Technology sharing during demand and supply planning in a networkbased supply chain environment  
Patent #
US 7,130,807 B1
Filed 11/22/1999

Current Assignee
Accenture Global Services Limited

Sponsoring Entity
Accenture LLP

Planning coordination systems for coordinating separate factory planning systems and a method of operation  
Patent #
US 5,983,194 A
Filed 02/25/1997

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
i2 Technologies Incorporated

System and method for displaying logistics information associated with a supply chain  
Patent #
US 6,486,899 B1
Filed 09/17/1999

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
i2 Technologies US Incorporated

Apparatus, system and method for reporting field replaceable unit replacement  
Patent #
US 6,684,180 B2
Filed 03/08/2001

Current Assignee
Google LLC

Sponsoring Entity
International Business Machines Corporation

Supply chain architecture  
Patent #
US 7,003,474 B2
Filed 08/15/2002

Current Assignee
iSuppli Inc.

Sponsoring Entity
iSuppli Inc.

Method of producing, selling, and distributing articles of manufacture  
Patent #
US 6,954,734 B1
Filed 07/20/2000

Current Assignee
LakeSouth Holdings LLC

Sponsoring Entity
World Factory Incorporated

System for facilitating the sale and shipment of cores  
Patent #
US 6,725,204 B1
Filed 07/12/2000

Current Assignee
Mark R. Gusley

Sponsoring Entity
Mark R. Gusley

Demand planning for configuretoorder and building blocksbased market environment  
Patent #
US 6,816,839 B1
Filed 05/04/2000

Current Assignee
International Business Machines Corporation

Sponsoring Entity
International Business Machines Corporation

Vehicle tracking, communication and fleet management system  
Patent #
US 6,611,755 B1
Filed 12/19/1999

Current Assignee
Trimble Navigation Limited

Sponsoring Entity
Trimble Navigation Limited

Communication schema of online system and method of ordering consumer product having specific configurations  
Patent #
US 6,609,108 B1
Filed 04/04/2000

Current Assignee
Ford Motor Company

Sponsoring Entity
Ford Motor Company

Method and system for the maximization of the range of coverage profiles in inventory management  
Patent #
US 6,341,266 B1
Filed 12/03/1998

Current Assignee
SAP SE

Sponsoring Entity
SAP SE

Method and system for managing a real time bill of materials  
Patent #
US 6,493,679 B1
Filed 05/26/1999

Current Assignee
Extreme Networks Inc.

Sponsoring Entity
Wireless Valley Communications Incorporated

System and method for forecasting intermittent demand  
Patent #
US 6,205,431 B1
Filed 10/29/1998

Current Assignee
Smart Software Inc.

Sponsoring Entity
Smart Software Inc.

Method and system for inventory management  
Patent #
US 6,061,691 A
Filed 12/18/1998

Current Assignee
iHeartMedia Inc.

Sponsoring Entity
B. Shane Fox, James A. Tiller Jr

Method for estimating stock levels in productiondistribution networks with inventory control  
Patent #
US 6,078,900 A
Filed 10/23/1998

Current Assignee
International Business Machines Corporation

Sponsoring Entity
International Business Machines Corporation

Method of and apparatus for production management  
Patent #
US 6,094,603 A
Filed 08/08/1996

Current Assignee
Fujifilm Corporation

Sponsoring Entity
Fuji Photo Film Co Limited

Decision support system for the management of an agile supply chain  
Patent #
US 6,151,582 A
Filed 02/24/1997

Current Assignee
US Philips Corporation

Sponsoring Entity
US Philips Corporation

Decision support system for the management of an agile supply chain  
Patent #
US 5,953,707 A
Filed 02/21/1997

Current Assignee
US Philips Corporation

Sponsoring Entity


Best can do matching of assets with demand in microelectronics manufacturing  
Patent #
US 5,971,585 A
Filed 09/09/1997

Current Assignee
GlobalFoundries Inc.

Sponsoring Entity
International Business Machines Corporation

Inventory management strategy evaluation system and method  
Patent #
US 5,963,919 A
Filed 12/23/1996

Current Assignee
Nortel Networks Limited

Sponsoring Entity
Nortel Networks Corporation

Method for part procurement in a production system with constrained resources  
Patent #
US 5,970,465 A
Filed 06/04/1997

Current Assignee
International Business Machines Corporation

Sponsoring Entity
International Business Machines Corporation

Method of estimating future replenishment requirements and inventory levels in physical distribution networks  
Patent #
US 6,006,196 A
Filed 05/01/1997

Current Assignee
Toshiba Global Commerce Solutions Holdings Corporation

Sponsoring Entity
International Business Machines Corporation

Method and apparatus for inventory control of a manufacturing or distribution process  
Patent #
US 5,819,232 A
Filed 03/22/1996

Current Assignee
DADE CHEMISTRY SYSTEMS INC.

Sponsoring Entity
DADE CHEMISTRY SYSTEMS INC.

Method for classifying sale amount characteristics, method for predicting sale volume, method for ordering for restocking, system for classifying sale amount characteristics and system for ordering for restocking  
Patent #
US 5,596,493 A
Filed 12/17/1992

Current Assignee
MEIJI MILK PRODUCTS CO. LTD. AND KAORU TONE

Sponsoring Entity
MEIJI MILK PRODUCTS CO. LTD. AND KAORU TONE

Point of supply use distribution process and apparatus  
Patent #
US 5,611,051 A
Filed 06/06/1995

Current Assignee
McKesson Information Solutions LLC

Sponsoring Entity
HBO COMPANY OF GEORGIA

Mailbox flag structure  
Patent #
US 5,445,317 A
Filed 03/04/1994

Current Assignee
Dennis W. Sokolowski

Sponsoring Entity
Dennis W. Sokolowski

Demand scheduled partial carrier load planning system for the transportation industry  
Patent #
US 5,265,006 A
Filed 12/26/1990

Current Assignee
Accenture Global Services Limited

Sponsoring Entity
ANDERSEN CONSULTING AN ILLINOIS PARTNERSHIP

24 Claims
 1. A computerimplemented system deploying parts in a distribution network, the system comprising:
a computer coupled with one or more databases, the computer configured to; access data associated with a plurality of parts at a plurality of locations in the distribution network; calculate a demand for each part at each location; estimate an availability leadtime for each part at each location; and generate a coverage function for each location based on the availability leadtime and the accessed data for each part.  View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
 9. A computerimplemented method for deploying parts in a distribution network, comprising:
accessing, by a computer, data associated with a plurality of parts at a plurality of locations in the distribution network; calculating, by the computer, a demand for each part at each location; estimating, by the computer, an availability leadtime for each part at each location; and generating, by the computer, a coverage function for each location based on the availability leadtime and the accessed data for each part.  View Dependent Claims (10, 11, 12, 13, 14, 15, 16)
 17. A computerreadable storage medium embodied with software of deploying parts in a distribution network, the software when executed using one or more computers is configured to:
access data associated with a plurality of parts at a plurality of locations in the distribution network; calculate a demand for each part at each location; estimate an availability leadtime for each part at each location; and generate a coverage function for each location network based on the availability leadtime and the accessed data for each part.  View Dependent Claims (18, 19, 20, 21, 22, 23, 24)
1 Specification
This application is a continuation of U.S. patent application Ser. No. 11/876,210, filed on 22 Oct. 2007 and entitled “OPTIMIZED DEPLOYMENT OF PARTS IN A DISTRIBUTION NETWORK”, which is a continuation of U.S. patent application Ser. No. 10/032,971, filed on 25 Oct. 2001 and entitled “OPTIMIZED DEPLOYMENT OF PARTS IN A DISTRIBUTION NETWORK”, now U.S. Pat. No. 7,337,031, and which claims the benefit under 35 U.S.C. §119(e) of U.S. Provisional Application Ser. No. 60/243,659, filed 26 Oct. 2000 and entitled “SYSTEM AND METHOD FOR OPTIMIZED DEPLOYMENT OF INVENTORY, OR REDISTRIBUTION OF EXISTING INVENTORY, ACROSS A MULTIECHELON DISTRIBUTION NETWORK” and which is related to U.S. patent application Ser. No. 10/033,103, filed 25 Oct. 2001 and entitled “REDISTRIBUTION OF PARTS IN A DISTRIBUTION NETWORK.”
U.S. patent application Ser. No. 11/876,210, U.S. patent application Ser. No. 10/033,103, U.S. Pat. No. 7,337,031, and U.S. Provisional Application Ser. No. 60/243,659 are commonly assigned to the assignee of the present application. The disclosure of related U.S. patent application Ser. No. 11/876,210, U.S. patent application Ser. No. 10/033,103, U.S. Pat. No. 7,337,031, and U.S. Provisional Application Ser. No. 60/243,659 are hereby incorporated by reference into the present disclosure as if fully set forth herein.
1. Technical Field of the Invention
This invention relates generally to the field of inventory distribution networks and more specifically to optimized deployment of parts in a distribution network.
2. Background of the Invention
Distribution networks may include one or more locations that receive parts from a vendor and distribute the parts within the distribution network in order to provide a customer with a product. The parts may be, for example, manufactured into a product within the distribution network. Distribution networks may include locations that both supply parts to and receive parts from other locations. Performance at each location is thus affected by the performance at its suppliers. As a result, maintaining an optimal inventory of parts at each location that best serves the customer while minimizing inventory costs poses a challenge for inventory managers.
In accordance with the present invention, disadvantages and problems associated with inventory deployment and redistribution techniques are reduced or eliminated.
According to one example of the present invention, a method for deploying parts is disclosed. Locations that include supply locations and demand locations are defined. A supply location supplies parts to a demand location. A demand is computed for each part at each location. An availability leadtime is estimated for each part at each location. A leadtime demand is computed for each part at each location using the availability leadtimes for the part. A stock level is computed for each part at each location. A completely filled demand is determined from the leadtime demands and the stock levels, and a partially filled demand is determined from the leadtime demands and the stock levels. A coverage function for the parts at the locations is generated from the completely filled demand and the partially filled demand.
Certain examples of the invention may provide one or more technical advantages. The present invention may be used to determine an optimized inventory deployment plan that describes the inventory at each location of a distribution network. The inventory deployment plan may optimize the ability of the distribution network to satisfy customer demand while conforming to business constraints. The inventory deployment plan may maximize the ability of the distribution network to fill orders, which may be calculated by minimizing the expected backorder of the distribution network. The present invention may be used to formulate a coverage function that is optimized to determine an optimized inventory deployment plan. Coverage may be used as a measure of customer service that describes the expected ability of each location to completely or partially fill a demand for a part. The present invention may be used to compute the expected number of partially and completely backordered demand for a part.
The present invention may be used to calculate a net demand for a part at a location that accounts for dependent demands and independent demands. A dependent demand at a location describes the parts that the location supplies to other locations in the distribution network, and an independent demand at a location describes the parts used at the location. Incorporating the independent and dependent demand into the demand may provide for a more accurate calculation of the demand. The present invention may be used to calculate a demand for a part at a location that takes into account the probability that the part is repaired and placed back into the inventory at the location. By taking into account the repaired parts, the calculation of the demand may be more accurate.
The present invention may be used to calculate the availability leadtime for a part at any number of supply locations. The demand location may order a certain proportion of parts from the supply locations in a particular order. The computation of the availability leadtime takes into account the probability that a supply location supplies the part, given that no other supply location has supplied the part, which may provide a more realistic calculation of availability leadtime. The replenishment leadtime at a demand end point may be computed as the availability leadtime at its supplier plus the transfer leadtime from the supplier to the demand end point.
The present invention may be used to calculate the expected number of backordered demand for a part at a location from the partially backordered and completely backordered demand. An equivalence relation between maximizing the coverage function and minimizing the sum of backorders may be determined.
Other technical advantages may be readily apparent to one skilled in the art from the figures, descriptions and claims included herein.
For a more complete understanding of the present invention and its features and advantages, reference is now made to the following description, taken in conjunction with the accompanying drawings, in which:
In one embodiment, locations 22 include a central location 22a and one or more warehouse locations 22bd. Although central location 22a and warehouse locations 22bd are illustrated, distribution network 20 may include any suitable number of central locations 22 and warehouse locations 22. Each location 22 may comprise a supply location and/or a demand location. A supply location supplies a part to a demand location, and may supply the part in response to an order for the part sent from the demand location. For example, warehouse location 22b supplies parts to warehouse location 22d, and warehouse locations 22bc supply parts to location 22d. A location 22 may comprise both a demand location and a supply location. For example, warehouse location 22b receives parts from central location 22a and supplies parts to warehouse location 22d. A supply endpoint such as central location 22a receives parts from one or more external supplies 24, for example, a vendor, and distributes the parts to warehouse locations 22bd. A demand endpoint such as warehouse location 22d provides parts to one or more external demands 32, for example, a customer.
Warehouse locations 22bd may include supply operations 26bd and/or repair operations 28bd. A supply operation 26 sends an order for a part to a supply location, which in response sends the part to supply operation 26. A repair operation 28 may receive a broken part from supply operation 26 and send the broken part to a repair center 30. Repair center 30 repairs the part and sends the repaired part to, for example, central location 22a or back to supply operation 26b. Alternatively, repair operation 28d may receive a broken part from supply operation 26d, repair the part, and send the repaired part back to supply operation 26d.
The inventory for each part at each location 22 is monitored, continuously or periodically. In response to the inventory falling below a predetermined level, an order is placed to bring the inventory position back up to a target level such as an optimized inventory level. A method for deploying and redistributing inventory of one or more parts among one or more locations to achieve optimized inventory levels is described in more detail with reference to
System 34 may include a computer system 35, a server 36, and a database 37, which may share data storage, communications, or other resources according to particular needs. Computer system 35 may include appropriate input devices, output devices, mass storage media, processors, memory, or other components for receiving, processing, storing, and communicating information according to the operation of system 34. As used in this document, the term “computer” is intended to encompass a personal computer, workstation, network computer, wireless data port, wireless telephone, personal digital assistant, one or more microprocessors within these or other devices, or any other suitable processing device.
Server 36 manages applications that generate optimized inventory deployment and redistribution plans. Server 36 includes one or more software components such as a preprocessing module 38 and a solver 39. Preprocessing module 38 manages input and output operations, and computes a net demand and a replenishment leadtime for each part at each location 22. Preprocessing module 38 computes a demand over leadtime, or leadtime demand, related to a number of parts in a pipeline. Preprocessing module 38 also generates mathematical formulations, which are transmitted to solver 39 for solving.
Preprocessing module 38 may include a deployment module 40 and a redistribution module. Deployment module 40 may be used to generate a coverage function that describes the distribution of parts among locations 22. Solver 39 optimizes the coverage function to determine an optimized distribution of parts. Solver 39 may comprise any suitable mathematical programming solver such as CPLEX by ILOG, INC. Redistribution module may be used to generate a transfer function that describes the cost of transferring parts among locations 22. Solver 39 optimizes the transfer function to determine an optimized manner of redistributing parts. As noted above, deployment may occur independently of redistribution. That is, inventory may be deployed, without ever being redistributed. Additionally, redistribution may redistribute parts according to an inventory plan generated by deployment module 40 or according to any suitable inventory plan.
Database 40 stores data that may be used by server 36. Data may include, for example, the history of the demand for each part at each location 22, the leadtime required to transport a part from one location 22 to another location 22, and the maximum space capacity at location 22. Computing system 35 and database 40 may be coupled to server 36 using one or more local area networks (LANs), metropolitan area networks (MANS), wide area networks (WANs), a global computer network such as the Internet, or any other appropriate wired, optical, wireless, or other links.
Processing module 38 initiates the method at step 46 by defining a number 1, 2, . . . , i, . . . , I of parts and a number 1, 2, . . . , j, . . . , J of locations 22. For example, j=1, 2, 3 and 4 and refer to warehouse locations 22ad, respectively. At step 48, data is accessed from database 37. Data may include, for example, a demand history of each part at each location 22. The demand history may describe the parts that each location 22 requires. Data may include repair history that may describe the capability of each location 22 to repair a part. Data may include the paths that may be used to transfer parts between locations 22, along with the costs associated with transporting parts along the paths. Data may include the cost of purchasing a part, the cost of storing a part in the location as a percentage of the purchase cost for the part, and a cost associated with ordering a part.
At step 50, a demand for each part at each location 22 is calculated. The demand may include a dependent demand and an independent demand. A dependent demand at location 22 describes the parts that location 22 supplies to other locations 22. An independent demand at location 22 describes parts used at location 22. The demand at location 22 may account for the probability that a part is repaired and placed back into the inventory at location 22. Demand may be calculated by starting at a demand endpoint and ending at a supply endpoint of distribution network 20. A method for calculating a demand for a part at each location 22 is described in more detail with reference to
A replenishment leadtime for each part at each location 22 is calculated at step 52. The replenishment leadtime for a part at location 22 describes the time required for location 22 to receive the part from another location 22. The replenishment leadtime may be computed by starting at a supply endpoint and ending at a demand endpoint. An availability leadtime for each part at each location 22 is estimated at step 54. The availability leadtime at a location 22 describes a waiting time due to back order at location 22 plus the transfer leadtime from a supplier to location 22 and the replenishment leadtime for the supplier of location 22. A method for estimating the availability leadtime of a part at location 22 is described in more detail with reference to
A coverage function is determined at step 56. The coverage function describes the expected ability of a location 22 to completely or partially fill an order for a part, and may be determined from the demand, availability leadtime of the part, and inventory level for the part at location 22. The coverage function may be described using the expected backorder of the part at location 22. A method for determining a coverage function is described in more detail with reference to
At step 62, redistribution module determines whether redistribution is required by calculating an excess and deficit for each part at each location 22 from the actual inventory and the optimal deployment. Redistribution of the inventory may be required if, for example, the actual inventory at each location 22 does not match the optimized inventory calculated for each location 22. As noted above, redistribution may redistribute parts according to the optimized inventory reported at step 61, or according to any suitable inventory plan. If redistribution is not required, deployment module 40 proceeds to step 63 to report any excess inventory and a recommendation to not perform a redistribution of parts. After reporting the result, the method is terminated. If redistribution is required, redistribution module proceeds to step 64 to check the transitions between locations 22. The transitions describe paths that may be used to transfer parts from one location 22 to another location 22.
At step 66, a transfer function describing the transfer of parts between locations 22 is optimized. Minimizing the total costs associated with transporting the parts may optimize the transfer function. A method for determining optimized transfer plans for transferring parts between locations 22 is described in more detail with reference to
At step 84, an independent and a dependent demand for part i at location j is determined. The independent demand for part i at location j may be represented by λ′_{ij}. The dependent demand for part i at location j may be represented λ_{ik}, where k is a demand end point for location j. At step 86, the repair capability r_{ij }for part i at location j is determined. The repair capability r_{ij }may be determined from the proportion of demand for part i at location j that is repairable at location j. Demand is calculated at step 88. Starting with demand end points j, demand λ_{ij }for part i is equal to its independent demand. For any location j that is not a demand end point, the demand λ_{ij }for part i may be calculated using Equation (1):
At step 92, deployment module 40 determines whether there is a next part for which a demand is to be determined. If there is a next part, deployment module 40 returns to step 80 to select the next part. If there is no next part, deployment module 40 proceeds to step 94 to output the calculated demand for each part at each location. After outputting the demand, the method is terminated.
At step 108, a probability P_{ilkj }of a supply location l_{k }filling an order for part i placed by a demand location j, given that the order is not filled by another supply location, is calculated. The probability P_{illj }for supply location l_{l }may be computed using Equation (2):
P_{illj}=α_{illj}C_{illj } (2)
At step 110, deployment module 40 determines whether there is a next supply location l_{k}. If there is a next supply location, deployment module 40 returns to step 106 to select the next supply location. The probability P_{ilkj }of the next supply location l_{k }filling an order for part i placed by demand location j, given that the order is not filled by another supply location, may be computed at step 108 using the process described by recursive Equations (3):
P_{il}_{k}_{j}=α_{il}_{k}_{j}C_{l}_{k}_{j }
where
C′_{il}_{l}_{j}=C_{il}_{l}_{j }
C′_{il}_{k}_{j}=C_{il}_{k}_{j}+(1−α_{il}_{k−1}_{j})C_{il}_{k−1}_{j }for k>1 (3)
If there is no next supply location at step 110, deployment module 40 proceeds to step 112 to output the probabilities of the supply locations l_{k }fulfilling an order placed by demand location j.
At step 113, an availability leadtime T_{ij }for each location j is calculated. Availability leadtime T_{ij }may be calculated according to recursive Equation (4):
The expected number of completely backordered demand B_{c }may be described by Equation (5):
where P(Xμ_{ij})=e^{−μ, }μ_{ij}^{x}/x! is the Poisson probability mass function for the distribution of demand with mean μ_{ij}, μ_{ik }represents the mean number of parts i in the pipeline at supply location k. The expected number of partially backordered demand B_{p }for part i at location j may be described by Equation (6):
The expected number of backorders EBO (S_{ij}) having the stock level S_{ij }of part i at location j may be defined using Equation (7):
At step 114, the replenishment leadtime Θ_{ij }for part i at demand location j is calculated. The replenishment leadtime Θ_{ij }for for part i at location j may be calculated using Equation (8):
Θ_{ij}=r_{ij}τ_{ij}+(1−r_{ij})T_{ij } (8)
where τ_{ij }represents the repair leadtime for part i at demand location j. The leadtime demand μ_{ij }of a part i at demand location j is estimated at step 116. The leadtime demand is related to a number of parts in a pipeline. The leadtime demand may be estimated using Equation (9):
μ_{ij}=λ_{ij}Θ_{ij } (9)
At step 118, deployment module 40 determines whether there is a next demand location. If there is a next demand location, deployment module 40 returns to step 104 to select the next demand location. If there is no next demand location, deployment module 40 proceeds to step 120 to determine whether there is a next part. If there is a next part, deployment module 40 returns to step 102 to select the next part. If there is no next part, deployment module 40 proceeds to step 122 to report the leadtime demand of each part at each location 22. After reporting the leadtime demand, the method is terminated.
At step 136, a completely filled demand D_{c }for part i at location j is calculated at step 136. A completely filled demand D_{c }may be described by Equation (10):
where P(Xμ_{ij})=e^{−μ}^{ij}μ_{ij}^{x}/x! is the Poisson probability mass function for the distribution of demand with mean μ_{ij}. A partially filled demand D_{p }for part i at location j is calculated at step 138. The partially filled D_{p }demand may be described by Equation (11):
where x is the percentage of partial fill allowed for a part. At step 140, a coverage function for part i at location j is determined. The coverage function for part i at location j describes the expected proportion filled demand for part i of location j, and may be expressed using Equation (12):
At step 142, deployment module 40 determines whether there is a next part. If there is a next part, deployment module 40 returns to step 134 to select the next part. If there is no next part, deployment module 40 proceeds to step 144 to determine whether there is a next location. If there is a next location, deployment module 40 returns to step 132 to select the next location. If there is no next location, deployment module 40 proceeds to step 146 to determine the coverage function for the number of parts at the number of locations. The coverage function may be expressed as the weighted average of coverage for the parts at locations. The coverage function for the parts at the locations may be expressed by Expression (13):
where β_{i }represents a weight of part i, which may be based on an importance measure of part i. At step 148, constraints for the coverage function may be defined. Constraints may include, for example, the following:
a. The weighted average of coverage for the parts at each location j is greater than or equal to the coverage target ω_{i }at location j, which may be expressed by Expression (13a):
b. The coverage for each part i at each location j is greater than or equal to the coverage target for part i at location j, which may be expressed by Expression (13b):
c. The number of new purchases for a part i, X_{ij }at location j, which may be expressed by Expression (13c):
X_{ij}=[S_{ij}+γ_{i}λ_{ij}−Y_{ij}]^{+}, ∀i, ∀j (13c)
where γ_{i }represents a proportion of a demand for a failed part i, and (x)^{+}=max(0,x).
d. The inventory investment at each location j is less than or equal to the inventory investment limit Inv_{j }for location j, which may be expressed by Expression (13d):
where C_{i }represents a purchase price for part i, and Y_{ij }represents an onhand inventory for part i at location j.
e. The overall inventory investment is less than or equal to an overall inventory investment limit Inv, which may be expressed by Expression (13e):
f. The overall inventory cost is less than or equal to overall budget B, which may be expressed by Expression (13f):
where h_{ij }represents a holding cost per unit of part i at location j, and k_{i }represents an order cost for part i.
g. The cost at each location j is less than or equal to a budget B_{j }at location j, which may be expressed by Expression (13g):
h. The total volume occupied by parts at each location j is less than or equal to the volume capacity limit V_{j }at location j, which may be expressed by Expression (13h):
i. The stock levels S_{ij }are integers, which may be expressed by Expression (13i):
S_{ij }are integers (13i)
At step 150, the coverage function is converted to a backorder function that corresponds to expected backorders, and the constraints are expressed in terms of backorders. Using the backorder function may provide for a simpler optimization process. The backorder function may be Expression (14):
Minimizing the backorder function is equivalent to maximizing the coverage function. The constraints may be expressed by Expressions (14a):
where S_{ij}^{M }represents the minimum stock level that guarantees the minimum target coverage. The high degree of nonlinearity of the constraints may be reduced by replacing the minimum target coverage constraint for each part at each location expressed by Expression (14a) with an equivalent constraint expressed as S_{ij}≧S_{ij}^{M}, ∀i, ∀j. The definition of the number of new purchases expressed by Expression (14b) may be replaced with a relaxed constraint expressed as X_{ij}≧S_{ij}+┌γ_{i}λ_{ij}−Y_{ij}┐, ∀i.
Maximization of the coverage function may be shown to be equivalent to minimizing the backorder function. The weighted average of coverage for distribution network 20 may be expressed by Expression (14b):
where θ_{ij}(S_{ij}) represents the coverage for part i at location j defined by Equation (14c):
The backorder function describes the total expected number of backorders for distribution network 20 and may be expressed by Expression (14d):
The following proposition may be established:
Proposition 1: Vector S_{j}═(S_{lj}, . . . , S_{lj}) satisfies the performance constraint expressed by Expression (14e):
if and only if vector S_{j }satisfies an expected backorders constraint expressed by Expression (14f):
A relationship between expected backorder EBO_{ij}(S_{ij}) and performance θ_{ij}(S_{ij}) may be established according to Equations (14g):
For any set of S_{ij }that satisfies the performance constraint expressed by Expression (14f), the following Expressions (14h) may be shown:
The stock level S_{ij }in question satisfies the location performance constraint expressed by Equations (14g). The steps may be reversed to prove the converse.
The following proposition describing the relationship between performance and coverage may be established:
Proposition: Maximizing systemwide coverage is equivalent to minimizing the total systemwide backorders.
The proposition may be established according to Equations (14i):
At step 154, the objective function that measures expected backorder as expressed by Expression (13) may be linearized. To linearize the objective function and constraints, the nonlinear terms of the objective function and constraints may be approximated by linear terms. The nonlinear terms are discrete and convex, so a firstorder linear approximation using the finite difference for two neighboring discontinuous points may be used to approximate each nonlinear term. Each nonlinear term in the objective function and the constraints is replaced with a continuous variable t_{ij }and a linearization constraint that describes the under estimation at points of discontinuity is added to the constraints.
The linearized objective function may be expressed by Expression (15):
The linearization constraint may be expressed by Expression (16):
t_{ij}≧m_{ij}(X−X_{ij})+b_{ij}, ∀S_{ij}<X_{ij}≦S_{upper}, ∀i,j (16)
where m_{ij}=P(X>X_{ij}μ_{ij}), b_{ij}=P(X>X_{ij}μ_{ij})(X−X_{ij})+EBO_{ij}(X_{ij}+1), and S_{upper }is the upper bound on the inventory for part i at location j. Other constraints may be expressed by Expressions (16a):
After linearizing, deployment module 40 solves the resulting mixed integerprogramming problem and determines the optimal stock levels for each part at each location.
An objective function measuring the systemwide total cost may also be defined at step 154. The total cost function may be expressed by Expression (16b):
The constraints for the total cost function may be expressed by Expressions (16c):
The total cost function and the constraints may also be linearized at step 154 in order to allow the objective function to be optimized by solver 39. The total cost objective function, as expressed by Expression (16b), may be linearized according to Expression (16d):
The constraints may be linearized according to Expressions (16e):
After linearizing, deployment module 40 solves the resulting mixed integerprogramming problem and determines the optimal stock levels for each part at each location.
Although an example of the invention and its advantages are described in detail, a person skilled in the art could make various alterations, additions, and omissions without departing from the spirit and scope of the present invention as defined by the appended claims.