GDonTaylor-Oct_2010 - WLU

Download Report

Transcript GDonTaylor-Oct_2010 - WLU

April 5, 2016
Alternative Dispatching Methods in
Truckload Trucking
Wilfrid Laurier University
October 1, 2010
Presentation
Overview
Motivation
Historical Dispatching
Various Alternative
Modes of Dispatch
The Latest (and Greatest)
Methods
Future Directions
Problem Domain
 US Logistics costs are approximately
$1.2 trillion annually, or 9.5% of GDP!
 6% of GDP is spent on transportation
costs alone. Approximately 80% of this
cost is in trucking.
 Freight availability is highly stochastic.
 Competitive environment results in
commodity pricing and prevents preplanning.
How would you dispatch?
Driver
Domicile
A
B
 Traditional
dispatch
minimizes
empty
repositioning
miles
Recent Research: Alternative
Dispatching Methods in TL Trucking
Minimization of empty
repositioning miles is good only
for the carrier
Excessive tour length &
retention statistics in LTL vs.
OTR
Quality of driver life is poor
Alternative dispatching methods
are greatly needed
Freight Imbalance
is Also a Big Problem
Best
Driver Domiciles
Domicile_finder
Output
Research in Truckload Trucking
Tour Length Reduction Strategies
Hub & Spoke Systems (Like LTL &
Intermodal)
Failure of H&S led to a focus on partitioned
fleets; but what are the effects on remaining
OTR?
Zone & lane dispatching
Optimal & dedicated models
‘Popcorn’ compromises
Regional dispatch models
‘Pipeline’ methods
Extended regional models
Hub & Spoke
Dispatching
A National Zone
Network
Picket Fence
Zone 3
Zone 1
Zone 5
Zone 4
Zone 2
Truckload Lanes
Optimal Seeking
Tour Building Routines
 More Difficult But Better Solutions
 Integer Programming Formulation
 Maximize:
 Subject to:
 Subject to:
 Subject to:
 Subject to:
 Subject to:
Loaded Miles Minus Empty Miles
Can Only Use Known Freight for
Loaded Moves
Drivers Begin and End at Domicile
All Transshipment Nodes Maintain
Balance Throughout
Maximum Number of Moves Per Tour
Sign Restrictions and Integrality
Dedicated Solutions
Optimal dispatch solutions easier with
trucks dedicated to a single customer;
perhaps with other freight mixed in
CL2
J21
CL3
J22
‘Popcorn’
Dispatching
Limited Random Dispatch
Seeks Middle Ground Between
Regularization and Randomness
Domicile Returns a Function of Limited
Network
Types of Dispatch Permitted:
Unlimited and Fully Random
Constrained by Moves or Miles
Heavily Constrained (Star Dispatch)
Pipeline Dispatching
Very Similar to
Intermodal with Rail
Requires 3 Dispatches
Regional Fleet
Development
5
13
11
9
12
1
4
8
41% of fleet
14
7
3
2
6
10
Integrated Regional Model with
Lane/Pipeline/OTR
5
13
11
9
12
1
4
8
14
7
3
2
6
10
Experimental Plan
• Baseline
• Integration of regions and lanes
• Regional dispatch given priority for
various service area sizes
• Lane dispatch given priority for
various service area sizes
• Lane/region operating flexibility
• Lane/region/pipeline/OTR
17
Performance of
Extended Regional
Models
• BR-IL-300 model utilizes 78 % of
all drivers with little negative effect
on remaining OTR fleet
• BL1-IR-300 utilizes 84% of all
drivers but neutralizes regions
• BR-IL-300-Flex and BR-IL-BP-300
not strong
18
SIMNET II Model
• Tracks/updates all drivers,
equipment, & loads
• Transient period to seed system
• Comprehensive & realistic
dispatch algorithm
• Metrics of interest to all parties;
drivers, carriers & shippers
19
Next Steps
• Combine real-time revenue
management with dispatch in the
truckload trucking industry
20
Q&A Period
21