Professional Documents
Culture Documents
= e
= e
s c
e e e
1
1
1
,
,
,
{0,1},
( , ) ( )
,
In the TSP formulation if we remove the third constraint set we
have the simple assignment problem, which can be easily
solved.
The addition of the third constraint set, commonly called sub-
tour elimination constraints, makes this a very difficult problem
to solve.
Questions about the TSP
Given a problem with n nodes, how many distinct
feasible tours exist?
How many arcs will the network have?
How many x
ij
variables will we have?
How could we quantify the number of subtour
elimination constraints?
The complexity of the TSP has led to several heuristic
or approximate methods for finding good feasible
solutions. The simplest solution we might think of is
that of the nearest neighbor.
Vehicle Routing: TSP, inventory routing, and vehicle routing
Traveling Salesman Problem (TSP): salesman visits n cities at
minimum cost
vehicle routing problem (VRP): m vehicles with capacity to
deliver to n customers who have volume requirement, time
windows, etc.
Inventory Routing: m vehicle to delivery to n customer with
time windows, vehicle and storage capacity constraints, and un-
specificed amount to be delivered.
Heuristics
1. Load points closest together on the same truck
2. Build routes starting with points farther from depot first
3. Fill the largest vehicle to capacity first
4. Routes should not cross
5. Form teardrop pattern routes.
6. Plan pickups during deliveries, not after all deliveries have
been made.
Illustration of VRP
(Outlier)
Depot
50
76
39
112
88
29
123
44
58
90
77
89
57
115
124
59 176
65
98 125
Truck Capacity = 250
What is the minimum # of trucks we would need? Maximum?
Vehicle Routing
Find best vehicle route(s) to serve a set of orders
from customers.
Best route may be
minimum cost,
minimum distance, or
minimum travel time.
Orders may be
Delivery from depot to customer.
Pickup at customer and return to depot.
Pickup at one place and deliver to another
place.
Complications
Multiple vehicle types.
Multiple vehicle capacities.
Weight, Cubic feet, Floor space, Value.
Many Costs:
Fixed charge.
Variable costs per loaded mile & per empty mile.
Waiting time; Layover time.
Cost per stop (handling).
Loading and unloading cost.
Priorities for customers or orders.
Pure Pickup or Delivery Problems.
Mixed Pickups and Deliveries.
Pickup-Delivery Problems.
Backhauls
More Complications
Time windows for pickup and delivery.
Hard vs. soft
Compatibility
Vehicles and customers.
Vehicles and orders.
Order types.
Drivers and vehicles.
Driver rules (DOT)
Max drive duration = 10 hrs. before 8 hr. break.
Max work duration = 15 hrs. before 8 hr break.
Max trip duration = 144 hrs.
Simple Models
Homogeneous vehicles.
One capacity (weight or volume).
Minimize distance.
No time windows or one time window per
customer.
No compatibility constraints.
No DOT rules.
VRP Solutions
Heuristics
Construction: build a feasible route.
Improvement: improve a feasible route.
Not necessarily optimal, but fast.
Performance depends on problem.
Worst case performance may be very poor.
Exact algorithms
Integer programming.
Branch and bound.
Optimal, but usually slow and applicable for small size
problem
Difficult to include complications.
The VRP is applicable in many practical situations directly related to
the physical delivery of goods such as
distribution of petroleum products,
distribution of industrial gases,
newspaper deliveries,
delivery of goods to retail store,
garbage collection and disposal,
package pick-up and delivery,
milk pick-up and delivery, etc.
the non-movement of goods such as
picking up of students by school buses,
routing of salesmen,
reading of electric meters,
preventive maintenance inspection tours,
employee pick-up and drop-off , etc.
APPLICATIONS OF VRP
A DSS
Employee Bus Routing
Commodity Distribution
In COVERS
Efficient Heuristic Procedures
NNH
MNNH
MSCWH
Simulation Features
Manipulate the System Generated Routes
Completely User Generated Routes
COVERS Handles
Multi-Depot VRP
Heterogeneous VRP
COVERS- COMPUTERIZED VEHICLE ROUTING SYSTEM
EMPLOYEE PICKUP VEHICLE ROUTING PROBLEM (EPVRP)
BANGALORE, KARNATAKA, INDIA
Indian Telephone Industries [ITI] Limited
Bharat Electronics Limited [BEL]
Hindustan Machine Tools [HMT]
Hindustan Aeronautics Limited [HAL]
Indian Space Research Organization [ISRO]
National Aeronautical Laboratory [NAL]
Central Machine Tools of India [CMTI]
AS A PROBLEM IN OR, A SIMPLIFIED EPVRP CAN BE DESCRIBED AS FOLLOWS:
GIVEN
A set (fixed number) of pick-up or delivery points,
The demand at every pick-up or delivery points (deterministic),
A set (fixed number) of vehicles (homogeneous) and
All relevant distance information across pick-up points.
IT IS REQUIRED TO FIND AN EFFECTIVE/EFFICIENT SOLUTION FOR
Assigning pick-up points to vehicles and
Sequencing pick-up points on the route of each vehicle
SO AS TO ACHIEVE THE OBJECTIVE OF
Minimizing the total distance traveled by the vehicles and/or the number of vehicles
used.
UNDER THE CONSTRAINTS THAT
Every route originates and terminates at the depot
The capacity of vehicle is restricted
The maximum distance (time) allowed for a vehicle on any route is within a pre-
specified limit
Each pick-up point is visited once only
Etc.,
AN ILP FORMULATION - EPVRP
Source : WATERS (1998)
ASSUMPTIONS
Vehicle capacity is known and constant (homogenous)
The number of vehicles available is known (at least the minimum
number of vehicles required is known)
The demand at every pick-up point is known (deterministic)
Maximum distance to be traveled by each vehicle is known and
constant for all vehicles
Demand at every pick-up point is less than or equal to vehicle
capacity
Every pick-up point is served by only one vehicle
Further, keeping in line with Waters formulation, the model formulation is
oriented towards routing during drop-back rather than pick-up. It is assumed
that the reverse logic holds good for pick-up.
Expanding the Scope of Linear Programming Solutions for Vehicle
Scheduling Problems. OMEGA, 16(6), 577-583
COMPUTATIONAL COMPLEXITY - OPTIMAL SOLUTION
#
PUP
Tot
Quantities
(Units)
# Variables
Including (0, 1)
Variables
# (0, 1)
Variables
#
Constraints
Optimal
Distance
(Km.)
# Routes # Iterations
(LINDO)
CPU Time
(AT 486)
4 61 48 16 60 13.2 1 45 2
5 71 75 25 85 26.4 2 330 3
6 79 108 36 114 28.6 2 353 6
7 106 147 49 147 31.0 2 2780 23
8 117 192 64 187 31.0 2 70724 80
9 132 243 81 225 37.4 2 43021 667
(11 Mts)
10 137 300 100 270 47.8 3 4963340 100800
(28 Hrs.)
Sutcliffe and Board (1990)
#
estimated that a simple extrapolation of Waters (1988) ILP approach using
the SCICONIC software might take nearly 1,20,000 years of CPU time on a VAX 8600 machine to solve a
VRP with 38 pick-up points!
#
Optimal Solution of VRP: Transporting Mentally Handicapped Adults to an Adult Training Center. JORS, 41(1), 61-67.
Nearest Insertion Heuristic (NIH)
Cheapest Insertion Heuristic (CIH)
Parallel Version of Clarke & Wright Heuristic (PCWH)
Sequential Version of Clarke & Wright Heuristic (SCWH)
Convex Hull Heuristic (CHH)
Nearest Neighbour Heuristic (NHH)
Modified NNH (MNNH)
Modified SCWH 1 (MSCWH-1)
Modified SCWH 2 (MSCWH-2)
HEURISTIC ALGORITHMS
CASE STUDY : DETAILS OF ROUTES, DISTANCES & SEAT UTILIZATION
Shift Timings #
Commuters
# Pickup
Points
#
Routes
Total Distance
per Trip (Km.)
Seat
Utilization (%)
A 06.15 02.15 PM 3659 303 64 1977.0 89.0
FG 07.30 04.15 PM 3999 313 66 2163.0 94.3
AG 08.45 05.30 PM 3042 286 53 1808.3 90.0
B 02.15 10.15 PM 975 242 30 1056.7 54.0
C
10.15 06.15 AM 40 ---- ---- ---- ----
Total
11715 410
213+
(426)
7005.0
(14010)
----
Ignored in our study
Each Bus Route (Trip) Repeated; Two Trips a day, Once for Pick-up and once for Drop-off.
Distinct Pick-up Points
COMPARATIVE PERFORMANCE (CASE STUDY) TOTAL DISTANCE
Procedures Shift 1
A
Shift 2
FG
Shift 3
AG
Shift 4
B
Total Distance
(Km.)
Savings
(in %)
CPU Time
PC/AT 486
@ 33 MHz
(Minutes)
Existing
Practice
(Manual)
1977.0 2163.0 1808.3 1056.7 7005.0 ----- ----
NIH 1875.8 2047.7 1734.1 890.3 6547.9 6.5 12
CIH 2155.2 2322.3 1914.2 1020.7 7412.4 - 5.8 52
PCWH 1803.5 2026.1 1761.1 1080.9 6671.6 4.76 19
SCWH 2139.2 2306.6 1889.2 1014.5 7349.5 - 4.9 18
CHH 1903.8 2047.7 1749.2 964.7 6665.4 4.85 55
NNH 1822.9 2063.2 1708.0 900.0 6494.1 7.29 1
MNNH 1817.7 2040.8 1740.7 858.9 6458.1 7.81 1
MSCWH-1 1796.2 2066.4 1687.5 910.2 6460.3 7.78 2
MSCWH-2 1799.4 2047.0 1688.5 908.5 6443.4 8.02 2
(Figures in Table represent travel distance in Km. For Pick-up only)
COMPARATIVE PERFORMANCE (CASE STUDY) TOTAL NUMBER ROUTES
Procedures Shift 1
A
Shift 2
FG
Shift 3
AG
Shift 4
B
Total Routes Reduction in
Trips (%)
Existing
Practice
(Manual)
64 66 53 30 213 -----
NIH 60 63 51 23 197 7.51
CIH 65 69 52 27 213 0
PCWH 63 68 56 36 223 - 4.7
SCWH 65 70 55 28 218 - 2.3
CHH 60 62 51 25 198 7.04
NNH 57 64 50 24 195 8.45
MNNH 57 63 51 23 194 8.92
MSCWH-1 58 63 49 24 195 8.45
MSCWH-2 58 63 49 24 194 8.92
Figures in Table represent number of trips for Pick-up only
Nearest Neighbour Heuristic (NHH)
Modified NNH (MNNH)
Modified SCWH-2 (MSCWH-2)
HEURISTIC ALGORITHMS - DSS IMPLEMENTATION
A Schematic Diagram of COVERS
DATA MANAGEMENT MODULE
General file
Depot Data File
Vehicle Data File
Pickup point Demand Data File
Inter-Stop Distance Data File
MODEL MANAGEMENT MODULE
Heuristic Procedures
Simulation Model
REPORT MANAGEMENT MODULE
Details of Route Sequence
Summary of Routes
Overall Summary of Routes
Depot wise Route Allocation
Vehicle Type wise Route Allocation
CONTROL MODULE
COMPUTER SYSTEM
USER