DECISION SUPPORT SYSTEM FOR REAL-TIME URBAN FREIGHT MANAGEMENT Hanna Grzybowska and Jaume Barceló Dept.

Slides:



Advertisements
Similar presentations
Traffic assignment.
Advertisements

CITY LOGISTICS “Is the process of totally optimising the logistics and transport activities by private companies in urban areas while considering the traffic.
An Exact Algorithm for the Vehicle Routing Problem with Backhauls
Gianluca Mastrolia DEEI - University of Trieste Via A. Valerio, Trieste - Italy 1 A B2B Application for a Freight.
Modeling Rich Vehicle Routing Problems TIEJ601 Postgraduate Seminar Tuukka Puranen October 19 th 2009.
Vehicle Routing & Scheduling: Part 1
Vehicle Routing & Scheduling Multiple Routes Construction Heuristics –Sweep –Nearest Neighbor, Nearest Insertion, Savings –Cluster Methods Improvement.
MANETs Routing Dr. Raad S. Al-Qassas Department of Computer Science PSUT
GEOG 111 & 211A Transportation Planning Traffic Assignment.
1 Sensor Relocation in Mobile Sensor Networks Guiling Wang, Guohong Cao, Tom La Porta, and Wensheng Zhang Department of Computer Science & Engineering.
Arc-based formulations for coordinated drayage problems Christopher Neuman Karen Smilowitz Athanasios Ziliaskopoulos INFORMS San Jose November 18, 2002.
Beneficial Caching in Mobile Ad Hoc Networks Bin Tang, Samir Das, Himanshu Gupta Computer Science Department Stony Brook University.
presented by Zümbül Bulut
Vehicle Routing & Scheduling
Ant Colonies As Logistic Processes Optimizers
EE 685 presentation Optimization Flow Control, I: Basic Algorithm and Convergence By Steven Low and David Lapsley Asynchronous Distributed Algorithm Proof.
An Inventory-Location Model: Formulation, Solution Algorithm and Computational Results Mark S. Daskin, Collete R. Coullard and Zuo-Jun Max Shen presented.
Norman W. Garrick Trip Assignment Trip assignment is the forth step of the FOUR STEP process It is used to determining how much traffic will use each link.
TRIP ASSIGNMENT.
Carl Bro a|s - Route 2000 Solving real life vehicle routing problems Carl Bro a|s International consulting engineering company 2100 employees worldwide.
Airline Schedule Optimization (Fleet Assignment II) Saba Neyshabouri.
Quadratic Programming Model for Optimizing Demand-responsive Transit Timetables Huimin Niu Professor and Dean of Traffic and Transportation School Lanzhou.
Decision for the location of Intermodal terminals in a rail-road network Anupam Kulshreshtha IIM - Lucknow.
A Dynamic Messenger Problem Jan Fábry University of Economics Prague
MATE: MPLS Adaptive Traffic Engineering Anwar Elwalid, et. al. IEEE INFOCOM 2001.
Package Transportation Scheduling Albert Lee Robert Z. Lee.
Institute of Production and Logistics – University of Natural Resources and Life Sciences, Vienna A Real-life Application of a Multi Depot.
Toshihide IBARAKI Mikio KUBO Tomoyasu MASUDA Takeaki UNO Mutsunori YAGIURA Effective Local Search Algorithms for the Vehicle Routing Problem with General.
Distributed Quality-of-Service Routing of Best Constrained Shortest Paths. Abdelhamid MELLOUK, Said HOCEINI, Farid BAGUENINE, Mustapha CHEURFA Computers.
Platzhalter für Bild, Bild auf Titelfolie hinter das Logo einsetzen Ann Melissa Campbell, Jan Fabian Ehmke 2013 Service Management and Science Forum Decision.
Deployment & Available-to-promise (ATP) in SCM EGN 5623 Enterprise Systems Optimization (Professional MSEM) Fall, 2011.
Technology and Society The DynamIT project Dynamic information services and anonymous travel time registration VIKING Workshop København Per J.
Routing and Scheduling in Transportation. Vehicle Routing Problem Determining the best routes or schedules for pickup/delivery of passengers or goods.
A Multi-Start Approach for Optimizing Routing Networks with Vehicle Loading Constraints Angel Juan Oscar Domínguez Department.
Network Models Tran Van Hoai Faculty of Computer Science & Engineering HCMC University of Technology Tran Van Hoai.
MIT ICAT ICATMIT M I T I n t e r n a t i o n a l C e n t e r f o r A i r T r a n s p o r t a t i o n Virtual Hubs: A Case Study Michelle Karow
ON SOME GRAPH RELATED PROBLEMS IN TRANSPORTATION ANALYSIS Jaume Barceló, Mª Paz Linares, Oriol Serch
Vehicle Routing & Scheduling: Developments & Applications in Urban Distribution Assoc. Prof. Russell G. Thompson Department of Infrastructure Engineering.
Regional Traffic Simulation/Assignment Model for Evaluation of Transit Performance and Asset Utilization April 22, 2003 Athanasios Ziliaskopoulos Elaine.
Transportation Planning, Transportation Demand Analysis Land Use-Transportation Interaction Transportation Planning Framework Transportation Demand Analysis.
On Activity-Based Network Design Problems JEE EUN (JAMIE) KANG, JOSEPH Y. J. CHOW, AND WILL W. RECKER 20 TH INTERNATIONAL SYMPOSIUM ON TRANSPORTATION AND.
+ Simulation Design. + Types event-advance and unit-time advance. Both these designs are event-based but utilize different ways of advancing the time.
The Application of The Improved Hybrid Ant Colony Algorithm in Vehicle Routing Optimization Problem International Conference on Future Computer and Communication,
Chapter 17 Scheduling. Management 3620Chapter 17 Schedule17-2 Overview of Production Planning Hierarchy Capacity Planning 1. Facility size 2. Equipment.
Mobile Agent Migration Problem Yingyue Xu. Energy efficiency requirement of sensor networks Mobile agent computing paradigm Data fusion, distributed processing.
“LOGISTICS MODELS” Andrés Weintraub P
EE 685 presentation Optimization Flow Control, I: Basic Algorithm and Convergence By Steven Low and David Lapsley.
Copyright 2003 Lynn Frock & Company. All Rights Reserved. 1 Five Ways to Build a Microsoft Project Schedule Lynn Frock, PMP Phone
Vehicle Routing & Scheduling
Experimental Evaluation of Real-Time Information Services in Transit Systems from the Perspective of Users Antonio Mauttone Operations Research Department,
Transportation Logistics CEE 498B/599I Professor Goodchild 4/18/07.
Content caching and scheduling in wireless networks with elastic and inelastic traffic Group-VI 09CS CS CS30020 Performance Modelling in Computer.
Log Truck Scheduling Problem
IT Applications for Decision Making. Operations Research Initiated in England during the world war II Make scientifically based decisions regarding the.
1 An Arc-Path Model for OSPF Weight Setting Problem Dr.Jeffery Kennington Anusha Madhavan.
Spring Routing: Part I Section 4.2 Outline Algorithms Scalability.
1 Travel Times from Mobile Sensors Ram Rajagopal, Raffi Sevlian and Pravin Varaiya University of California, Berkeley Singapore Road Traffic Control TexPoint.
Copyright © 2014 by McGraw-Hill Education (Asia). All rights reserved. 13 Aggregate Planning.
Dynamically Computing Fastest Paths for Intelligent Transportation Systems - ADITI BHAUMICK ab3585.
Urban Planning Group Implementation of a Model of Dynamic Activity- Travel Rescheduling Decisions: An Agent-Based Micro-Simulation Framework Theo Arentze,
Management Science 461 Lecture 7 – Routing (TSP) October 28, 2008.
Distance Vector Routing
September 2008What’s coming in Aimsun: New features and model developments 1 Hybrid Mesoscopic-Microscopic Traffic Simulation Framework Alex Torday, Jordi.
Tabu Search Applications Outlines: 1.Application of Tabu Search 2.Our Project with Tabu Search: EACIIT analytics.
Tuesday, March 19 The Network Simplex Method for Solving the Minimum Cost Flow Problem Handouts: Lecture Notes Warning: there is a lot to the network.
A MapReduced Based Hybrid Genetic Algorithm Using Island Approach for Solving Large Scale Time Dependent Vehicle Routing Problem Rohit Kondekar BT08CSE053.
T-Share: A Large-Scale Dynamic Taxi Ridesharing Service
Copyrights (H.Rashidi & E.Tsang)
Planning the transportation of elderly to a daycare center
Chapter 6 Network Flow Models.
Presentation transcript:

DECISION SUPPORT SYSTEM FOR REAL-TIME URBAN FREIGHT MANAGEMENT Hanna Grzybowska and Jaume Barceló Dept. of Statistics and Operations Research CENIT (Center for Innovation in Transport) Universitat Politècnica de Catalunya

CONCEPTUAL APPROACHES TO REAL TIME FLEET MANAGEMENT AND THE ROLE IF ICT TECHNOLOGIES 2 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

Supplier Supplier’s Supplier Manufacturer Wholesaler/Distributor Retailer CITY AREA - client CITY LOGISTICS SCENARIO 3 Fleet management in urban areas has to explicitly account for the dynamics of traffic conditions leading to congestions and variability in travel times affecting the distribution of goods and the provision of services City Logistics activities are impacted by traffic congestion  must consider time-varying traffic congestion and operational constraints in routing and logistics optimization models Last-mile logistics Decisions must take into account all factors conditioning the problem  Decision Support System (DSS) The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

4 GPS Satellites Vehicle’s Position Vehicle Vehicle’s Data Updated Route Cellular Antenna Vehicle’s Data Global Positioning System (GPS) GPS device pickups signal from satellites GPS device calculates position Establish communication with network Vehicle Data is sent to Fleet Management Center. Fleet Manager updates routes and returns them to vehicle. Fleet management Centre ICT TECHNOLOGIES AND REAL-TIME FLEET MANAGEMENT The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

BARCELONA’S TRAFFIC INFORMATION SYSTEM 5 Current and Short-Term Forecasted Travel Time The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

6 Initial Demand and Fleet Specifications ROUTING AND SCHEDULING MODULE Initial Operational Plan DYNAMIC ROUTER AND SCHEDULER Dynamic Operational Plan CONCEPTUAL SCHEME FOR REAL-TIME FLEET MANAGEMENT SYSTEMS (Adapted from Regan, Jaillet, Mahmassani) Known/predicted demand for service Known/predicted driver/vehicle availability Load Acceptance Policies Pool of accepted demands REAL-TIME INFORMATION New demands  Unsatisfied demands  Traffic conditions  Fleet availability ASSUMING: A given Initial Operational Plan Fleets equipped with AVL tecnologies (i.e. GPS+GPRS) A Real-Time Information System OBJECTIVE: Design a DSS for Dynamic Routing and Scheduling The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

New Customer Request Cancelation of Service Arrival of Vehicle at Location Changes in Travel Times Delays in Delivery Start Times Breakdown of Vehicle Changes in Time Windows Bounds DYNAMIC ROUTER AND SCHEDULER Insertion Heuristics Local Search Operators TABU SEARCH Reactive Strategies One-by-one Pooling Preventive Strategies Vehicle Relocation Waiting Strategies Drive-First Wait-First Combined DYNAMIC EVENT AND VEHICLE TRACKING SIMULATOR NEW ROUTING PLAN EXTERNAL EVENTS INTERNAL EVENTS INTERNAL EVENTS DYNAMIC MONITORING SOLVING STRATEGIES PROPOSED DECISION SUPPORT SYSTEM FOR REAL-TIME FLEET MANAGEMENT The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

DYNAMIC ROUTER AND SCHEDULER Decide where to insert the new customer. New customer arrives at time t > 0. Fleet vehicles can be in one of the three status: –In service at some customer i (SER). –Moving to the next planned customer on the route or waiting at the customer location to start service within the time window. (MOV). –Idle at the Depot, without a previously assigned route (IDL). –Waiting at the client’s i (WAIT) This status determines when a vehicle should be diverted from its current route, be assigned to a new one if is idle or keep the planned trip. Whenever a new customer arrives, the status of a vehicle must be known to compute travel times for this new customer. If the vehicle has a MOV status, the travel time is computed from the current position of the vehicle to the location of the new customer. If the vehicle has IDL status, the travel time is just the travel time from the depot to the new customer. If the vehicle is has SER status, the amount of time needed to arrive to the customer is the remaining service time at the current customer plus the travel time between the current customer and the new customer. If the vehicle has WAIT status it can be send to other client. The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

1.Initial fleet scheduling 2.Start services while tracking vehicles 3.New client call is received at time t 4.Check vehicle positions and current travel times 5.Reject unfeasible routes (insertion / diversion) 6.Recalculate routes. 7.Execute new plan. In Transit In Service New Client DYNAMIC ROUTING AND SCHEDULING: Example of Dynamic Insertion Heuristic The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

DYNAMIC REROUTING WITH REAL-TIME INFORMATION (ADJUSTMENT SOF ROUTE 2 - GREEN AND 3 - BLUE ) 10 section removed from the original route new section of modified original route new client request registered at time t initially known client already visited by the assigned vehicle (before time instant t) initially known client still not visited by the assigned vehicle (before time instant t) on of the fleet vehicle serving route 2 at time t The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

AVERAGE LINK TRAVEL TIMES AND VARIANCES (Barcelona “Eixample”-CBD) 11 Working exclusively with average travel times may lead to significant deviations in city logistics problems where temporality is an important factor such as the VRPTW. A report on urban distribution in Barcelona (Robusté, 2005) found that: There were more than 62,000 commercial outlets  more than 60,000 daily unloading operations Service time between 13 and 16 minutes 50% of deliveries were made by 11:00 hrs The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

DEALING WITH TIME-DEPENDENT TRAVEL TIMES d i = departure time from client i s i = sevice time for client i T ij (d i ) = travel time from i to j when departing at time d i from client i T ij (d i )  T ij (d i ’) 12 Ad Hoc version of the algorithm of Ziliaskopoulos and Mahmassani Calculates the time-dependent shortest paths from all nodes of to one, specified as destination point. The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

REAL-TIME TRAFFIC INFORMATION SYSTEM REAL-TIME FLEET MONITORING SYSTEM The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

FRAMEWORK 14 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

TIME-DEPENDENT TRAVEL TIMES DATA FORECASTING MODULE MULTIPLE SIMULATION ITERATIONS “future” travel times actual travel times historical travel times data base SINGLE SIMULATION ITERATION t0t0 TRAFFIC SIMULATOR cjH(t0)cjH(t0) c j H (t 0 +1) cjH(t0+T)cjH(t0+T) … cjP(t0)cjP(t0)c j P (t 0 +1)cjP(t0+T)cjP(t0+T) … t0t0 Historical Data Base Actual Travel Times + Expected Travel Times ATIS 15 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

Time-Dependent Routing Plan: sequences of clients to visit Time-Dependent Shortest Paths: sequences of nodes between the clients for the vehicle to go through Present Travel Times Data Base Vehicle Fleet Performance Simulator Vehicles’ current: status position load List of clients: served omitted INPUT OUTPUT e.g.: instant of appearance of new event Depending on the current vehicle’s status, the value of time left to complete: service waiting trip on the current arc EXTERNAL TRIGGER INTERNAL EVENT e.g.: arrival to a client whose TW is closed end of providing a client with service VEHICLE FLEET PERFORMANCE SIMULATOR 16 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

MATHEMATICAL FORMULATION OF THE VEHICLE ROUTING PROBLEM WITH PICKUP AND DELIVERIES AND TIME WINDOWS Subject to : Each customer is served by the same vehicle Time windows feasibility Precedence requirement Route for vehicle k, from o(k) to d(k) Route and vehicle loads requirements 17 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

PICK UP & DELIVERY PROBLEMS WITH TIME WINDOWS 18 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

COMPOSITE HEURISTICS TO CONTRUCT THE INITIAL AND THE DYNAMIC ROUTING AND SCHEDULING 19 ALGORITHM PROVIDING THE INITIAL SOLUTION Based on the Simple Pairing Approach Based on the Sweep Algorithm Based on Customers’ Aggregation Areas PARALLEL TABU SEARCH PROCEDURE USING SIMULTANEOUSLY TWO LOCAL SEARCH HEURISTICS Pickup and Delivery Pair Shift Operator Pickup and Delivery Pair Exchange Operator POST-OPTIMIZATION Pickup and Delivery Pair Rearrange Operator 2-opt procedure The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

P2 D1 P6 D6 DEPOT D7 P7 P3 D3 P5 D5 D4 D2 P4 P1 NEIGHBOURHOOD 1NEIGHBOURHOOD 2 NEIGHBOURHOOD 3 20 Algorithm Initial Solution 1.Order known clients (Client’s Sorting Pre-Processing Algorithm using the Definitions of Customers’ Aggregation Areas 2.Create initial solution Select the farthest client in the listing Find its PD partner and delete them both from the listing IF it is the first iteration OR it is not possible to insert the pair into existing route THEN create route: depot-pickup customer-delivery customer-depot ELSE insert the pair in location causing minimal increment of the cost of the existing route The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

PARALLEL TABU SEARCH Works with two concurrent different search processes: The Pickup and Delivery Pair Shift Operator (NPDPSO) The Pickup and Delivery Pair Exchange Operator (NPDPEO). 21 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

POST-OPTIMIZATION PROCESS Post-optimization is realised by: Pickup and Delivery Pair Shift Operator 22 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

COMPUTATIONAL EXPERIMENTS AND RESULTS 23 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

THE SELECTED SCENARIO: Barcelona’s CBD 24 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

KTS seminar at Linkoping University, Norrkoping Clients Depot Downtown area of Barcelona. Commercial activities and tourism. It covers 747 hectares 1,570 links y 721 nodes. THE SIMULATION MODEL 25 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

100 Customers: Constant demand and service time. Time windows 1 Depot; Opening hours 08:00 – 16:00 Fleet: 8 homogeneous vehicles with large capacity. Vehicles are equipped with GPS and real-time communication system with fleet manager. Real-time traffic information system. Simulation time: 10 hours (07:00 – 17:00) MODELING ASSUMPTIONS 26 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

Table 1. Collection of the DSS performance testing scenarios TESTING SCENARIOS ScenarioN o Initially Known ClientsTDSP Calculator TriggeredTraffic InformationUsed Module 1100once at the startaverage staticIRSM 2100at the start and after multiple timestime-dependantIRSM+DRSM with CRRM 380at the start and after multiple timestime-dependantIRSM+DRSM with CRRM Initial solutionFinal solution Scenario Ini N o r Total Travel Time [s] Total Waiting Time [s] Solution Cost [s] Fin N o r Total Travel Time [s] Total Waiting Time [s] Solution Cost [s] Service Level % % % Table 2. Results on fleet performance depending on the traffic information input 27 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

TESTING SCENARIOS, USING SPI 28 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

TESTING SCENARIOS USING CSR 29 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

SERVICE LEVELS IN SCENARIOS WITH DYNAMIC CLIENTS 30 N o stat cust. (%) service level The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

TOTAL TRAVEL TIMES IN SCENARIOS WITH DYNAMIC CLIENTS 31 total travel time [s] N o stat cust. (%) The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

CONCLUSIONS The obtained results prove that the performance of the fleet strongly depends on the traffic information used to create and update the routing and scheduling plan. The usage of the time-dependent shortest paths, computed whenever a new even occurs, brings better results than when the average travel times’ estimates are employed, even in the case when not all the information on the clients to be served is initially available. The comparison of the initially planned and performed routing and scheduling plan indicates that the total travel time, total waiting time and the solution cost are always higher when executed. It is due to the fact that the plan does not take into consideration the most recent and forecasted changes in the traffic flow. No customer is left unserved. 32 The bi-weekly logistics and routing research group meeting, NICTA, 29th July 2011

33 THANK YOU VERY MUCH FOR YOUR ATTENTION