Routing shipments according to criticality

0Associated
Cases 
0Associated
Defendants 
0Accused
Products 
38Forward
Citations 
0
Petitions 
10
Assignments
First Claim
1. A computerimplemented system for routing shipments according to criticality, the system comprising one or more software components collectively operable to:
 access an initial solution to an optimization problem of routing a plurality of predetermined shipments to a plurality of predetermined locations using a plurality of vehicles, the initial solution comprising a plurality of loads such that each shipment is routed within exactly one load and a global cost across all the loads is minimized, the initial solution being generated independent of the criticality of the shipments;
insert, into each of one or more critical loads in a current solution, one or more noncritical shipments that are within a neighborhood of the critical load, a critical load being a load containing at least one critical shipment; and
execute one or more local search operations to improve the initial solution, the executed local search operations comprising at least one of;
a first operation comprising splitting each of one or more selected critical loads in a current solution into two new critical loads;
a second operation comprising, for each of one or more selected pairs of critical loads in a current solution, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair; and
a third operation comprising, for each of one or more selected critical loads in a current solution that are indirect critical loads having at least one intransit stop, breaking up the indirect critical load into a plurality of new direct critical loads having no intransit stops and for each of one or more selected pairs of critical loads, each selected pair of critical loads including at least one new direct critical load, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair.
10 Assignments
0 Petitions
Accused Products
Abstract
A computerimplemented method for routing shipments according to criticality includes accessing an initial solution to an optimization problem of routing multiple shipments to multiple locations using multiple vehicles, the initial solution including multiple loads such that each shipment is routed within exactly one load and a global cost across all loads is minimized, the initial solution being generated independent of the criticality of the shipments. Into each of one or more critical loads in a current solution, one or more noncritical shipments are inserted that are within a neighborhood of the critical load, a critical load being a load containing at least one critical shipment. One or more local search operations are executed to improve the initial solution, the operations including at least one of: (a) splitting each of one or more selected critical loads in a current solution into two new critical loads; (b) for each of one or more selected critical load pairs in a current solution, move a sequence of stops from one critical load in the pair to the other critical load in the pair and/or swap two sequences of stops between the critical loads in the pair; and (c) for each of one or more selected critical loads in a current solution that are indirect critical loads having at least one intransit stop, break up the indirect critical load into a plurality of new direct critical loads having no intransit stops and execute operation (b) on each of one or more selected critical load pairs, each selected critical load pair including at least one new direct critical load.
46 Citations
View as Search Results
System and method for optimization of group shipments to reduce shipping costs  
Patent #
US 7,937,330 B2
Filed 01/15/2009

Current Assignee
Deem Incorporated

Sponsoring Entity
Rearden Commerce Inc

Enhanced Location Information For Points of Interest  
Patent #
US 20110112761A1
Filed 09/09/2010

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

System and method for automatic review of travel changes and improved suggestions and rules set  
Patent #
US 7,966,213 B2
Filed 10/16/2006

Current Assignee
Deem Incorporated

Sponsoring Entity
Rearden Commerce Inc

ENHANCED LOCATION INFORMATION FOR POINTS OF INTEREST  
Patent #
US 20110172911A1
Filed 03/24/2011

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

ENHANCED LOCATION INFORMATION FOR POINTS OF INTEREST  
Patent #
US 20110172910A1
Filed 03/24/2011

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Method and system for testing of policies to determine cost savings  
Patent #
US 7,743,002 B2
Filed 09/30/2005

Current Assignee
Deem Incorporated

Sponsoring Entity
Rearden Commerce Inc

System for optimization of cost management  
Patent #
US 7,660,743 B1
Filed 10/15/2004

Current Assignee
Deem Incorporated

Sponsoring Entity
Rearden Commerce Inc

System and Method for Tracking Spending Based on Reservations and Payments  
Patent #
US 20090006142A1
Filed 06/26/2007

Current Assignee
Rearden Commerce Inc

Sponsoring Entity
Rearden Commerce Inc

System and method for optimization of group shipments to reduce shipping costs  
Patent #
US 7,496,520 B1
Filed 07/22/2005

Current Assignee
Deem Incorporated

Sponsoring Entity
Rearden Commerce Inc

System and Method for Optimization of Group Shipments to Reduce Shipping Costs  
Patent #
US 20090125355A1
Filed 01/15/2009

Current Assignee
Deem Incorporated

Sponsoring Entity
Deem Incorporated

System and method for flexible handling of rules and regulations in labor hiring  
Patent #
US 7,596,566 B1
Filed 02/24/2005

Current Assignee
Deem Incorporated

Sponsoring Entity
Talaris Corporation

System, computerimplemented method, and software for vessel scheduling for product distribution  
Patent #
US 20080010357A1
Filed 06/23/2006

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
JDA Software Group Incorporated

Navigation system  
Patent #
US 20060129314A1
Filed 12/09/2005

Current Assignee
LG Electronics Inc.

Sponsoring Entity
LG Electronics Inc.

Method and system for testing of policies to determine cost savings  
Patent #
US 20060190314A1
Filed 09/30/2005

Current Assignee
Deem Incorporated

Sponsoring Entity
Deem Incorporated

Method and systems for personal restaurant assistant  
Patent #
US 8,126,776 B2
Filed 06/30/2006

Current Assignee
Deem Incorporated

Sponsoring Entity
Rearden Commerce Inc

System, computerimplemented method, and software for vessel scheduling for product distribution  
Patent #
US 8,321,354 B2
Filed 06/23/2006

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
JDA Software Group Incorporated

System and method for constraintbased reduction of a solution space for vehicle routing  
Patent #
US 8,386,397 B1
Filed 06/30/2010

Current Assignee
Amazon Technologies

Sponsoring Entity
Amazon Technologies

Enhanced location information for points of interest  
Patent #
US 8,655,487 B2
Filed 03/24/2011

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Enhanced location information for points of interest  
Patent #
US 8,725,407 B2
Filed 09/09/2010

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Enhanced location information for points of interest  
Patent #
US 8,751,153 B2
Filed 01/11/2013

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Enhanced location information for points of interest  
Patent #
US 8,849,563 B2
Filed 03/24/2011

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

System and method of vessel scheduling for product distribution  
Patent #
US 8,972,304 B2
Filed 10/22/2012

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
JDA Software Group Incorporated

Enhanced location information for points of interest  
Patent #
US 8,972,165 B2
Filed 01/27/2014

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

MULTILEVEL DISTRIBUTION PLANNING  
Patent #
US 20150168947A1
Filed 12/13/2013

Current Assignee
Oracle International Corporation

Sponsoring Entity
Oracle International Corporation

Enhanced location information for points of interest  
Patent #
US 9,082,100 B2
Filed 01/06/2014

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Cost model analysis and breakdown for cost buildup  
Patent #
US 9,161,994 B1
Filed 03/29/2005

Current Assignee
Talaris Corporation

Sponsoring Entity
Deem Incorporated

Apparatus and method to provide community pricing  
Patent #
US 9,226,975 B1
Filed 12/30/2004

Current Assignee
Deem Incorporated

Sponsoring Entity
Deem Incorporated

System and method of vessel scheduling for product distribution  
Patent #
US 9,665,845 B2
Filed 03/02/2015

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
JDA Software Group Incorporated

Multilevel distribution planning  
Patent #
US 9,704,125 B2
Filed 12/13/2013

Current Assignee
Oracle International Corporation

Sponsoring Entity
Oracle International Corporation

Methods for delivering a parcel to a restricted access area  
Patent #
US 9,928,749 B2
Filed 04/28/2017

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Unmanned aerial vehicle including a removable power source  
Patent #
US 9,957,048 B2
Filed 04/28/2017

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Unmanned aerial vehicle pickup and delivery systems  
Patent #
US 9,969,495 B2
Filed 04/28/2017

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Unmanned aerial vehicle including a removable parcel carrier  
Patent #
US 9,981,745 B2
Filed 04/28/2017

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Methods for picking up a parcel via an unmanned aerial vehicle  
Patent #
US 10,202,192 B2
Filed 01/12/2018

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Unmanned aerial vehicle and landing system  
Patent #
US 10,453,022 B2
Filed 01/12/2018

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Delivery vehicle including an unmanned aerial vehicle support mechanism  
Patent #
US 10,460,281 B2
Filed 01/12/2018

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

Unmanned aerial vehicle chassis  
Patent #
US 10,482,414 B2
Filed 01/12/2018

Current Assignee
United Parcel Service Of America Incorporated

Sponsoring Entity
United Parcel Service Of America Incorporated

System and method of vessel scheduling for product distribution  
Patent #
US 10,535,033 B2
Filed 05/26/2017

Current Assignee
JDA Software Group Incorporated

Sponsoring Entity
JDA Software Group Incorporated

Router utility for a parcel shipping system  
Patent #
US 6,571,213 B1
Filed 12/30/1999

Current Assignee
PINEY BOWES INC.

Sponsoring Entity
PINEY BOWES INC.

Order allocation to select from inventory locations stocking few units of inventory  
Patent #
US 6,622,127 B1
Filed 05/10/2000

Current Assignee
JUNE RAY LIMITED

Sponsoring Entity
Kaiser Foundation Hospitals

Bill of lading transmission and processing system for less than a load carriers  
Patent #
US 6,401,078 B1
Filed 04/03/2000

Current Assignee
RL Carriers Inc.

Sponsoring Entity
RL Carriers Inc.

Dynamic traffic based routing algorithm  
Patent #
US 6,285,951 B1
Filed 06/30/2000

Current Assignee
Murata Machinery Limited

Sponsoring Entity
PRI AUTOMATION INC.

Allocation method for generating a production schedule  
Patent #
US 5,548,518 A
Filed 05/31/1994

Current Assignee
International Business Machines Corporation

Sponsoring Entity
International Business Machines Corporation

Method for warehousing and delivery of articles  
Patent #
US 5,403,147 A
Filed 06/25/1992

Current Assignee
KAO Corporation

Sponsoring Entity
KAO Corporation

Imagebased document processing system providing for priority document shipment  
Patent #
US 5,237,158 A
Filed 10/10/1989

Current Assignee
Unisys Corporation

Sponsoring Entity
Unisys Corporation

Job scheduling system  
Patent #
US 5,093,794 A
Filed 08/22/1989

Current Assignee
United Technologies Corporation

Sponsoring Entity
United Technologies Corporation

35 Claims
 1. A computerimplemented system for routing shipments according to criticality, the system comprising one or more software components collectively operable to:
access an initial solution to an optimization problem of routing a plurality of predetermined shipments to a plurality of predetermined locations using a plurality of vehicles, the initial solution comprising a plurality of loads such that each shipment is routed within exactly one load and a global cost across all the loads is minimized, the initial solution being generated independent of the criticality of the shipments; insert, into each of one or more critical loads in a current solution, one or more noncritical shipments that are within a neighborhood of the critical load, a critical load being a load containing at least one critical shipment; and execute one or more local search operations to improve the initial solution, the executed local search operations comprising at least one of; a first operation comprising splitting each of one or more selected critical loads in a current solution into two new critical loads; a second operation comprising, for each of one or more selected pairs of critical loads in a current solution, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair; and a third operation comprising, for each of one or more selected critical loads in a current solution that are indirect critical loads having at least one intransit stop, breaking up the indirect critical load into a plurality of new direct critical loads having no intransit stops and for each of one or more selected pairs of critical loads, each selected pair of critical loads including at least one new direct critical load, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair.  View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13)
 14. A computerimplemented method for routing shipments according to criticality, comprising:
accessing an initial solution to an optimization problem of routing a plurality of predetermined shipments to a plurality of predetermined locations using a plurality of vehicles, the initial solution comprising a plurality of loads such that each shipment is routed within exactly one load and a global cost across all the loads is minimized, the initial solution being generated independent of the criticality of the shipments; inserting, into each of one or more critical loads in a current solution, one or more noncritical shipments that are within a neighborhood of the critical load, a critical load being a load containing at least one critical shipment; and executing one or more local search operations to improve the initial solution, the executed local search operations comprising at least one of; a first operation comprising splitting each of one or more selected critical loads in a current solution into two new critical loads; a second operation comprising, for each of one or more selected pairs of critical loads in a current solution, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair; and a third operation comprising, for each of one or more selected critical loads in a current solution that are indirect critical loads having at least one intransit stop, breaking up the indirect critical load into a plurality of new direct critical loads having no intransit stops and for each of one or more selected pairs of critical loads, each selected pair of critical loads including at least one new direct critical load, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair.  View Dependent Claims (15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26)
 27. Software for routing shipments according to criticality, the software being embodied in computerreadable media and when executed operable to:
access an initial solution to an optimization problem of routing a plurality of predetermined shipments to a plurality of predetermined locations using a plurality of vehicles, the initial solution comprising a plurality of loads such that each shipment is routed within exactly one load and a global cost across all the loads is minimized, the initial solution being generated independent of the criticality of the shipments; insert, into each of one or more critical loads in a current solution, one or more noncritical shipments that are within a neighborhood of the critical load, a critical load being a load containing at least one critical shipment; and execute one or more local search operations to improve the initial solution, the executed local search operations comprising at least one of; a first operation comprising splitting each of one or more selected critical loads in a current solution into two new critical loads; a second operation comprising, for each of one or more selected pairs of critical loads in a current solution, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair; and a third operation comprising, for each of one or more selected critical loads in a current solution that are indirect critical loads having at least one intransit stop, breaking up the indirect critical load into a plurality of new direct critical loads having no intransit stops and for each of one or more selected pairs of critical loads, each selected pair of critical loads including at least one new direct critical load, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair.  View Dependent Claims (29, 30, 31, 32, 33, 34, 35)
 28. A computerimplemented system for routing shipments according to criticality, comprising:
means for accessing an initial solution to an optimization problem of routing a plurality of predetermined shipments to a plurality of predetermined locations using a plurality of vehicles, the initial solution comprising a plurality of loads such that each shipment is routed within exactly one load and a global cost across all the loads is minimized, the initial solution being generated independent of the criticality of the shipments; means for inserting, into each of one or more critical loads in a current solution, one or more noncritical shipments that are within a neighborhood of the critical load, a critical load being a load containing at least one critical shipment; and means for executing one or more local search operations to improve the initial solution, the executed local search operations comprising at least one of; a first operation comprising splitting each of one or more selected critical loads in a current solution into two new critical loads; a second operation comprising, for each of one or more selected pairs of critical loads in a current solution, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair; and a third operation comprising, for each of one or more selected critical loads in a current solution that are indirect critical loads having at least one intransit stop, breaking up the indirect critical load into a plurality of new direct critical loads having no intransit stops and for each of one or more selected pairs of critical loads, each selected pair of critical loads including at least one new direct critical load, performing at least one of moving a sequence of stops from one critical load in the pair to the other critical load in the pair and swapping two sequences of stops between the critical loads in the pair.
1 Specification
This invention relates in general to computerimplemented logistics, and more particularly to routing shipments according to criticality.
Routing of shipments between businesses and other locations is a significant problem in commercial environments. The problem of optimally routing shipments can be constructed as the wellknown Vehicle Routing Problem (VRP), which is an important optimization problem in operations research. VRPs are an extension of the wellknown Travelling Salesman Problem (TSP). In general, VRPs are described as the problem of delivering multiple shipments to multiple locations using multiple vehicles. In solving VRPs, the goals is to define a set of routes subject to capacity, time, or other suitable constraints such that each shipment is delivered in exactly one of the routes while the global cost of these routes is minimized. VRPs belong to a class of difficult problems, referred to as NPhard problems, for which no efficient solution algorithms are known.
Certain exact algorithms and heuristic procedures have been proposed for solving VRPs. Due to computer processing power and other limitations, the exact algorithms can only be used to solve relatively small VRPs. Exact algorithms also often lack flexibility. Local search heuristics have been effective for solving certain largescale realworld problems. The general framework for a local search includes two phases: (1) route construction and (2) route improvement. In route construction, initial routes are constructed using various heuristics, such as insertion heuristics. In the improvement phase, local search operations are applied to initial routes to obtain better routes. The most commonly used operations are moving a sequence of stops from one route to another and swapping two sequences of stops between two routes. These operations belong to a class of operations referred to as kopt.
Although the previous models for solving VRPs have been useful for certain applications, many important practical issues have not been addressed. As a result, it remains important to extend VRP formulation according to real world transportation needs and to develop corresponding solution strategies.
According to the present invention, disadvantages and problems associated with prior techniques for routing shipments may be reduced or eliminated.
In one embodiment of the present invention, a method for routing shipments according to criticality includes accessing an initial solution to an optimization problem of routing multiple shipments to multiple locations using multiple vehicles, the initial solution including multiple loads such that each shipment is routed within exactly one load and a global cost across all loads is minimized, the initial solution being generated independent of the criticality of the shipments. Into each of one or more critical loads in a current solution, one or more noncritical shipments are inserted that are within a neighborhood of the critical load, a critical load being a load containing at least one critical shipment. One or more local search operations are executed to improve the initial solution, the operations including at least one of: (a) splitting each of one or more selected critical loads in a current solution into two new critical loads; (b) for each of one or more selected critical load pairs in a current solution, move a sequence of stops from one critical load in the pair to the other critical load in the pair and/or swap two sequences of stops between the critical loads in the pair; and (c) for each of one or more selected critical loads in a current solution that are indirect critical loads having at least one intransit stop, break up the indirect critical load into a plurality of new direct critical loads having no intransit stops and execute operation (b) on each of one or more selected critical load pairs, each selected critical load pair including at least one new direct critical load.
Particular embodiments of the present invention may provide one or more technical advantages. Particular embodiments may generate a solution to a modified VRP that takes the criticality of shipments into consideration, the solution preferably specifying desirable relatively full executable loads for a substantial number of critical shipments. In particular embodiments, the solution can be considered a first order approximate solution to a multiple time period VRP. Particular embodiments use a number of local search operations to identify desirable solutions to the problem. These local search operations may be relatively simple, flexible, and effective, yielding good solutions for largescale real world scenarios in a reasonable amount of time. One or more other technical advantages may be readily apparent to those skilled in the art from the figures, descriptions, and claims included herein.
To provide a more complete understanding of the present invention and the features and advantages thereof, reference is made to the following description taken in conjunction with the accompanying drawings, in which:
As an example of the problem addressed according to the present invention, consider the typical situation in which shippers solve VRPs periodically to execute transportation plans. In general, VRPs may be described as the problem of delivering multiple shipments to multiple locations using multiple vehicles. In general, a load (also referred to as a route) may be defined as a movement of one or more shipments between two or more stops. In solving VRPs, the goal is to define a set of loads subject to capacity, time, or other constraints such that each shipment is delivered in exactly one of the loads while the global cost of these loads is minimized. VRPs belong to a class of difficult problems, referred to as NPhard problems, for which no efficient solution algorithms have been developed. It is common for shippers to have shipments that do not need to be executed in a current time period. Thus, in addition to determining a set of loads containing these shipments, decisions have to be made concerning which loads will be executed in the current time period and which will instead be executed in future time periods. In full generality, the above problem is a complex multiple time period VRP, usually with forecasted shipments in future time periods.
A shipment may be called “critical” if it must be executed in the current time period, otherwise it may be called “future” (i.e. “noncritical”). A load may be called “critical” if it contains one or more critical shipments. A load may also be called “executable” if it is acceptable according to certain business rules, for example only and without limitation, more than forty thousand pounds and less than four stops in transit. The present invention considers an approximation to the multiple time period VRP described above by solving a VRP only for the current time period. Intuitively, it is desirable to build critical loads with significant consolidation, for example, full vehicles, straight routes, and few stops, because critical loads by definition must be executed. Moreover, it is also desirable to build as many as possible executable noncritical loads so that they can be considered for execution in the current time period, for example, to increase consolidation. Thus, in one embodiment, system 10 uses the following primary objectives: (1) minimize the unit cost for critical loads, then (2) maximize the total volume of executable loads. It may be desirable to execute noncritical executable loads, for example, to decrease inventory, increase visibility, and increase the time available to negotiate with carriers to transport these loads. The problem solved according to the present invention is a modified VRP incorporating criticality constraints as described herein.
In operation of system 10, in an initialization stage, optimizer 14 accesses the information representing the shipments to be routed and provides this information to solver 16 for generation of an initial solution to a basic VRP. The initial solution, for example, may include only critical shipments (thus excluding future shipments) or may include all shipments regardless of criticality. Significantly, because the initial solution is a solution to a basic VRP, the initial solution treats all the shipments within the problem space (i.e. either all critical shipments or all shipments) equally and does not distinguish among the shipments based on their criticality. Optimizer 14 receives the initial solution from solver 16. Before, substantially simultaneous with, or after causing the initial solution to be generated, optimizer 14 accesses the parameters optimizer 14 will use to execute one or more operations and accesses a strategy script or other suitable instructions that specify the order in which optimizer 14 will execute the operations. Optimizer 14 may receive this information from a user through user interface 12, retrieve the information from database 18, or access the information in any other appropriate manner.
In a construction stage, optimizer 14 may execute a Move Future Shipments operation that attempts to greedily fill each critical load with future shipments in an appropriate neighborhood of the critical load. The Move Future Shipments operation may or may not be executed, depending on the strategy optimizer 14 is pursuing, as specified in the strategy script or other suitable instructions. In one embodiment, it is desirable to execute the Move Future Shipments operation whether the initial solution is for a VRP that incorporates only critical shipments or for a VRP that incorporates all shipments. For example, if solver 16 generates an initial solution to a VRP that incorporates only critical shipments, then execution of the Move Future Shipments operation may be desirable to minimize global plan cost by filling light critical loads with noncritical shipments. If the initial solution incorporates all the shipments (without regard to criticality), then the Move Future Shipments operation may be desirable to minimize global plan cost by moving certain noncritical shipments from noncritical loads to light critical loads. The Move Future Shipments operation is described more fully below with reference to
In an improvement stage, optimizer 14 may execute one or more appropriate local search operations to improve the initial solution (possibly as modified by the Move Future Shipments operation). In essence, local search heuristics are based on trial and error. In general, given an instance (F, c) of an optimization problem, where F is some feasible set of solutions and c is a cost mapping, a local search algorithm involves choosing a neighborhood N(x), which maps a solution x to a subset of F:
N:F>2^{F}.
The neighborhood N(x) is searched at a solution xεF for improvements using the logic: improve(x) is true if there exists any yεN(x) such that c(y)<c(x), but is false otherwise. The general local search algorithm can be summarized:
 Begin
 x: some initial starting point in F;
 while improve(x)=true
 x=improve(x);
 return t;
 End
 Begin
As indicated above, the local search algorithm starts at some initial feasible solution xεF and uses the subroutine improve( ) to search for a better solution in the neighborhood of that solution. So long as an improved solution exists, the improved solution is adopted it and the neighborhood search is repeated for the new solution. When a local optimum is reached, the local search algorithm stops. To apply this approach to a particular problem, we must first decide how to obtain an initial feasible solution. Next, we must choose a “good” neighborhood for the problem at hand and a method for searching it. This choice is usually guided by intuition, because very little theory is available as a guide. The design of effective local search algorithms has been, and remains, very much an art.
In one embodiment, during the improvement stage, optimizer 14 may execute one or more of the following local search operations, without limitation: (1) a Split Critical Loads operation that, for a single execution, splits a critical load into two loads if there are a sufficient number of future shipments in the neighborhood of the critical load to create two executable loads; (2) an Optimize Critical Loads operation that, for a single execution, moves a sequence of stops from one load to another or swaps two sequences of stops between two loads (and may remove future shipments to make room for critical shipments or add future shipments to fill critical loads); and (3) a ReOptimize Critical Loads operation that breaks up a light critical load into a set of direct loads (i.e. having no intransit stops), then applies the Optimize Critical Loads operation to the new loads along with all the other critical loads. These local search operations are described more fully below with reference to
In a particular embodiment, according to the strategy script or other suitable instructions, optimizer 14 applies the Split Critical Loads operation to each critical load within the initial solution (preferably as modified by the Move Future Shipments operation). Each execution of the Split Critical Loads operation is for a single critical load. The operation is preferably applied to the critical loads in order of priority, such that in one embodiment the most important and fullest critical loads may be split first. Once the Split Critical Loads operation has been applied to all critical loads in the current best solution, optimizer 14 selects a new best solution. Still according to the strategy script or other instructions, optimizer 14 applies the Optimize Critical Loads operation to appropriate pairs of critical loads in the current best solution (selected after executing the Split Critical Loads operation). Each execution of the Optimize Critical Loads operation is for a single pair of critical loads. Once the Optimize Critical Loads operation has been applied to all appropriate pairs of critical loads in the current best solution, optimizer 14 selects a new best solution. Still according to the strategy script or other instructions, optimizer 14 next applies the ReOptimize Critical Loads operation to each critical load in the current best solution (selected after executing the Optimize Critical Loads operation). Each execution of the ReOptimize Critical Loads operation is for a single critical load. The operation may be applied to the critical loads in order of priority, such that the most important and fullest critical loads may be broken up first. Once the ReOptimize Critical Loads operation has been applied to all critical loads in the current best solution, optimizer 14 selects a new best solution. In this particular embodiment, this solution is deemed the solution to the modified VRP incorporating criticality constraints. Those skilled in the art will appreciate that, although a particular solution strategy has been described, the present invention encompasses any solution strategy falling within the scope of the claims appended hereto.
Before describing the operations in more detail, it may be useful to introduce the following terminology applicable in certain embodiments:
Insertion Percentage: Given a load, a shipment, and a location to insert the shipment, the insertion percentage is calculated as the percentage distance increase of the load as the result of inserting the shipment at the specified location.
Percent Full: Given a vehicle having a weight capacity of MaxWeight and a volume capacity of MaxVolume. The percent full of a shipment is the greater of the ratios weight/MaxWeight and volume/MaxVolume. Similarly, we can define percent full for a load as its percent full at its origin (i.e. before any shipments in the load are delivered to their destinations).
 Unit Cost: Unit cost is the cost per percent full per distance. For example, assume a load has two shipments and costs $300. Suppose also that each shipment is 50% full and one has the distance 100 miles, the other 200 miles. Then the cost per percent full per mile is 300/(0.5*100+0.5*200)=2.
Rating Engine: In practice, the cost of a load may depend on many factors, including mileage, number of stops, carrier, equipment, geographic region, etc. For simplicity, we assume there is an external rating engine (not explicitly shown) that is capable of returning the cost of a load to system 10 given the various parameters of the load as input.
Scheduling Engine: In addition, loads will be subject to various time window constraints. For example, shipments will usually have pickup and delivery dates, and locations will have business hours. For simplicity, we assume there is an external scheduling engine (not explicitly shown) that is capable of checking for scheduling violations for a load.
Neighborhood: The definition of a neighborhood of a load may be important for local searching. Though the present invention provides advantages when used in the context of general neighborhoods, it may be more desirable to choose a particular simple neighborhood for ease of presentation. Thus, assuming only one origin, the neighborhood for a load may be defined to include the set of loads/shipments whose destinations are within a MaxNeighborhoodDistance of the load'"'"'s destination.
Priority: As an extension of criticality, in one embodiment a priority may be associated with each noncritical (future) shipment, which preferably takes a nonnegative integer value. Among noncritical (future) shipments, some shipments are more important than others (meaning it is more important that these shipments be shipped sooner rather than later). The more important a shipment, the lower the integer value of its priority. The priority of a load may be defined as the highest priority of all its shipments.
Example execution of the Move Future Shipments operation begins at step 30, where optimizer 16 sorts all critical loads within the existing solution by priority, if any, and then within priority by decreasing percent full. This preferably allows the most important and fullest critical loads to be filled with shipments first. Optimizer 16 selects a first critical load at step 32, determines its neighborhood at step 34 as specified by the MaxNeighborhoodDistance parameter, and determines all the future shipments whose destinations are within MaxNeighborhoodDistance of the critical load'"'"'s destination and are not already in critical loads at step 36. At step 38, these future shipments are sorted by decreasing order of priority if any, then by increasing order of insertion percentage, then by decreasing order of percent full. Optimizer 16 selects a first future shipment at step 40 and inserts the future shipment into the critical load at step 42. This insertion is accepted if the resulting load is valid according to any applicable constraints and the insertion also does not increase the total cost of the plan, relative to the existing solution, by more than the value of the MaxCostIncreaseAllowed parameter. Optimizer 16 will either accept or reject this insertion at step 44.
If there is another future shipment to consider at step 46, then the execution returns to step 42 and iterates until all the appropriate future shipments have been considered. If there is another critical load to consider at step 48, then the execution returns to step 34 and iterates until all critical loads have been considered, at which point execution of the Move Future Shipments operation ends. Although the Move Future Shipments operation is described as involving each critical load and each of the future shipments determined at step 36, the present invention contemplates this operation being executed using any suitable subset of critical loads within an existing solution and any suitable subset of future shipments.
Example execution of the Split Critical Loads operation begins at step 60, where optimizer 16 may sort all the critical loads within the existing solution by priority, if any, and then within priority by decreasing percent full, as also described above with reference to the Move Future Shipments operation. This may allow the most important and fullest critical loads to be split first. Alternative embodiments may omit such sorting of critical loads. Optimizer 16 selects a first critical load at step 62. In a particular embodiment, the critical load must be multistop and have a percent full greater than MinPercentFullRequired. Suppose a load has N stops, in one embodiment only split suboperations that split the critical load into two loads with stops (1, 2, . . . , k) and (k+1, . . . , N) are considered. There are N−1 such split suboperations. For a split suboperation at stop k, the distance savings of the split suboperation may be calculated as:
distance savings=distance between stop k and stop k+1.
At step 64, optimizer 16 sorts all the N−1 possible split suboperations for the critical load by distance savings. Optimizer 16 may estimate whether it is possible for a split suboperation to create two executable loads by examining at all future shipments in the neighborhood for the critical load. Any split suboperations resulting in nonexecutable loads may be discarded.
Optimizer 16 selects the first split suboperation (preferably the one with the largest distance savings) at step 66, makes temporary changes to split the critical load into two new loads at step 68, and then uses the Move Future Shipments operation with parameters (MaxNeighborhoodDistance, MaxInsertionPercent, 0) to fill both new loads with future shipments at step 70. If both new loads have a percent full greater than MinPercentFullRequired and the unit cost across all the critical loads is improved, then all the temporary changes associated with the split suboperation are accepted. Otherwise, the temporary changes are discarded and the next split suboperation is attempted. Optimizer 16 either accepts or rejects the split suboperation at step 72.
If there is another split suboperation to consider at step 74, then the method returns to step 68 and iterates until all the appropriate split suboperations have been considered. If there is another critical load to consider at step 76, then the method returns to step 64 and iterates until all critical loads have been considered, at which point execution of the Split Critical Loads operation ends. Although described as involving each critical load and each split operation that was deemed acceptable in connection with step 64, the present invention contemplates execution of the Split Critical Loads operation for any appropriate subset of critical loads in an existing solution and any suitable subset of split suboperations for a critical load.
After the split, based on an appropriate MaxNeighborhoodDistance, 25 miles for example, the future shipment to D is used to fill load C and the future shipment to E is used to fill load AB. The resulting loads are shown in the lower diagram. Now we calculate the unit cost to determine whether the changes associated with the (AB, C) split should be accepted. For simplicity, assume a cost of $1 per mile. Distances for the shipments from the origin are d(A)=100, d(B)=130, d(C)=160, d(D)=180, and d(E)=140. The products of the distance and percent full for each shipment are p(A)=100*0.4=40, p(B)=130*0.3=39, p(C)=160*0.3=48, p(D)=180*0.6=108, p(E)=140*0.2=28. Thus, for the original critical load ABC, the unit cost is 180/(40+39+48)=1.42, and for the new loads, the unit cost is (140+180)/(40+39+48+108+28)=1.22. Since 1.22<1.42, improvement is made and the changes are accepted. In one embodiment, the split (A, BC) will not be attempted since changes have already been made (i.e. load ABC no longer exists to split).
In one embodiment, the parameters for the Optimize Critical Loads operation are MaxNeighborhoodDistance, MaxInsertionPercent, MaxDestinationDistance, MinPercentFullRequired, and MinDistanceSavings, without limitation. Parameters may be received from a user using user interface 12, retrieved from database 18, or accessed in any other suitable manner. As for the Split Critical Loads operation, the MinPercentFullRequired parameter is preferably used to define whether a load is executable, although one or more other or additional parameters may be used for this purpose if appropriate. The MinDistanceSavings parameter may be used to eliminate moves and swaps that do not appear sufficiently promising, of which there may be many. The MaxNeighborhoodDistance and MaxInsertionPercent parameters are, as described more fully below, used for the Move Future Shipments operation after a move or swap suboperation has been executed.
Example execution of the Optimize Critical Loads operation begins at step 80, where optimizer 16 selects a first pair of critical loads, referred to as loads A and B for simplicity. In one embodiment, due to the very large number of possible move and swap suboperations for each pair of critical loads, the search space is narrowed by discarding the pair of critical loads if their destination distance is greater than the MaxDestinationDistance parameter. Thus, if the destination distance for the pair is not acceptable according to the MaxDestinationDistance parameter at step 82, then the pair is discarded at step 84 and a next pair of critical loads may be selected (which may include one of the critical loads included in the previous pair). Otherwise, the optimizer 16 creates quadruples for all the possible move and swap suboperations between the critical loads of the pair at step 86. As with the Split Critical Loads operation, optimizer 16 may estimate whether it is possible for a suboperation to create two executable loads by examining all future shipments in the neighborhood for the critical loads in the pair. A quadruple for a suboperation resulting in nonexecutable loads may be discarded. If there is another pair of critical loads to be considered at step 88, then execution returns to step 82 and iterates until all suitable quadruples have been considered. If all pairs of critical loads have been considered at step 88, then at step 90 optimizer 16 sorts all the quadruples for all pairs by distance savings, which may be calculated as described above for the Split Critical Loads operation.
Optimizer 16 selects the first quadruple (preferably the quadruple with the largest distance savings) at step 92 and makes temporary changes to implement the suboperation on the pair to create two new loads at step 94. If capacity constraints are violated at step 96, then optimizer 16 removes all future shipments from the two new loads at step 98 and uses the Move Future Shipments operation to fill both new loads with future shipments at step 99. If capacity constraints are not violated at step 96, then optimizer 16 proceeds directly to step 99. If both of the new loads have a percent full more than MinPercentFullRequired and the unit cost across all critical loads is improved, then all the temporary changes associated with the suboperation are accepted at step 100. If not, the suboperation is rejected at step 100, temporary changes are discarded, and the next quadruple is attempted.
In one embodiment, if loads A and B are changed to loads A′ and B′ after one successful suboperation (accepted at step 100), optimizer 16 discards all remaining quadruples involving loads A or B at step 102, creates new pairs of critical loads involving loads A′ and/or B′ at step 104, and then returns to step 80 where execution continues in the manner described above. However, in this case steps 80 through 86 preferably concern only the new pairs and associated quadruples. That is, only new critical load pairs that involve loads A′ and/or B′ will be considered for selection at step 80 and, accordingly, quadruples are created only for the selected new critical load pair at step 86. Quadruples not involving loads A or B are unaffected by the suboperation accepted at step 100 and remain in the pool of all possible quadruples. The distance savings for each new quadruple is determined and, at step 90, all possible quadruples (including the new quadruples) are sorted by distance savings before the next quadruple is selected.
If the suboperation is not accepted at step 100 and there is another quadruple to consider at step 106, execution returns to step 94. If there is not another quadruple to consider at step 106, execution of the Optimize Critical Loads operation ends. While described as involving each pair of critical loads (those existing before the Optimize Critical Loads operation begins as well as those that are created as a result of successful suboperations accepted at step 100) and each quadruple that is deemed acceptable in connection with step 90, the present invention contemplates executing the Optimize Critical Loads operation for any suitable subset of pairs of critical loads and any suitable subset of quadruples for a pair of critical loads.
Now we determine whether an improvement has been made by comparing the unit cost of the first solution (A, BC) with the unit cost of the second solution (AB, C). First, the product of the distance and percent full is calculated for each shipment. For the shipments to A and to B this is 100*0.5=50 and for the shipment to C this is 150*0.5=75. For the first solution, there are two critical loads A and BC and the unit cost for the critical loads is:
250/(50+50+75)=1.43.
For the new solution, there is only one critical load AC the cost is 100+75=175 and the unit cost is:
175/(50+75)=1.4.
Since 1.4<1.43, the changes associated with the move suboperation are accepted, in fact yielding the optimal solution for this simple scenario. Note, however, that the total cost has increased from $250 to $275. This is not surprising since the objective for a VRP with criticality constraints according to the present invention is different from that of a basic VRP.
Example execution of the ReOptimize Critical Loads operation begins at step 110, where optimizer 16 may sort all the critical loads within the existing solution by priority, if any, and then within priority by decreasing percent full, as also described above with reference to the Move Future Shipments operation and the Split Critical Loads operation. This may allow the most important and fullest critical loads to be broken up first. Alternative embodiments may omit such sorting of critical loads. Optimizer 16 selects a first critical load at step 112. In one embodiment, the critical load must be multistop and have percent full greater than MinPercentFullRequired. At step 114, optimizer 16 breaks up the critical load into a set of direct critical loads, denoted for simplicity by S. At step 115, for each direct critical load contained in set S, the optimizer 16 uses the direct critical load as a seed to create quadruples with all the critical loads not contained in set S. The ReOptimize Critical Loads operation then proceeds as for the Optimize Critical Loads operation described above.
Therefore, at step 116, optimizer 16 sorts the quadruples by distance savings, which may be calculated as described above. As with the Split Critical Loads and Optimize Critical Loads operations, optimizer 16 may estimate whether it is possible to create two executable loads by examining all future shipments in the neighborhood for the critical loads in the pair associated with a quadruple. A quadruple for a suboperation resulting in nonexecutable loads may be discarded. Optimizer 16 selects the first quadruple (preferably the quadruple with the largest distance savings) at step 118 and makes temporary changes to implement the suboperation on the pair to create two new loads at step 120. If capacity constraints are violated at step 122, optimizer 16 removes all future shipments from the two new loads at step 124 and uses the Move Future Shipments operation to fill both these new loads with future shipments at step 125. If capacity constraints are not violated at step 122, optimizer 16 proceeds directly to step 125. If both new loads have a percent full more than MinPercentFullRequired and the unit cost across all the critical loads is improved, then all the temporary changes associated with the suboperation are accepted at step 126. If not, the suboperation is rejected at step 126, the temporary changes are discarded, and the next quadruple is attempted.
In one embodiment, if loads A and B are changed to loads A′ and B′ after one successful suboperation, optimizer 16 discards all remaining quadruples involving A or B at step 128 and returns to step 110, where the critical loads (including the new direct critical loads created as a result of the suboperation accepted at step 126) are again sorted. If the suboperation is not accepted at step 126 and there is another quadruple to consider at step 130, then execution returns to step 120. If there is not another quadruple to consider at step 130, then execution of the ReOptimize Critical Loads operation ends. Although described as involving each critical load (both those existing before the ReOptimize Critical Loads operation begins and those created as a result of successful suboperations accepted at step 126), and each quadruple that is deemed acceptable in connection with step 116, the present invention contemplates executing the ReOptimize Critical Loads operation for any suitable subset of critical loads and any suitable subset of quadruples for a critical load.
During the construction stage, at step 208, optimizer 14 may execute the Move Future Shipments operation to fill each critical load in the initial solution with future shipments that are in an appropriate neighborhood of the critical load. During an improvement stage, at steps 210–220, optimizer 14 may execute one or more local search operations to improve the initial solution (preferably as modified according to execution of the Move Future Shipments operation). In the illustrated embodiment, optimizer 14 applies the Split Critical Loads operation to each critical load within the initial solution at step 210, selects a new best solution at step 212, applies the Optimize Critical Loads operation to the appropriate pairs of critical loads in the current best solution at step 214, selects a new best solution at step 216, applies the ReOptimize Critical Loads operation to each critical load in the current best solution at step 218, and selects a new best solution at step 220. In this particular embodiment, the new best solution selected at step 220 is deemed the solution to the modified VRP incorporating criticality constraints at step 222, and the method ends. Those skilled in the art will appreciate that, although a particular solution strategy has been described, the present invention encompasses any solution strategy falling within the scope of the claims appended hereto.
Although the present invention has been described with several embodiments, a plethora of changes, substitutions, variations, alterations, and modifications may be suggested to one skilled in the art, and it is intended that the invention encompass all such changes, substitutions, variations, alterations, and modifications as fall within the spirit and scope of the appended claims.