Presentation is loading. Please wait.

Presentation is loading. Please wait.

Multi Agent Approach for Time Based Vehicle Arrangement in Reputed Taxi Companies. 6037PS2013016 – D. A. P. Peiris Supervisors Mr. D.D.A. Gamini Mr. B.

Similar presentations


Presentation on theme: "Multi Agent Approach for Time Based Vehicle Arrangement in Reputed Taxi Companies. 6037PS2013016 – D. A. P. Peiris Supervisors Mr. D.D.A. Gamini Mr. B."— Presentation transcript:

1 Multi Agent Approach for Time Based Vehicle Arrangement in Reputed Taxi Companies.
6037PS – D. A. P. Peiris Supervisors Mr. D.D.A. Gamini Mr. B. Hettige

2 Introduction Taxi service is supporting to national transportation system to reduce their work load and make comfortable transport experience to customers. Last few years, new software architecture for managing the taxi service at the tactical and operational level has emerged. It views the taxi service as composed of a set of intelligent software agents, each responsible for one or more activities in the taxi service and each interacting with other agents in the planning and execution of their responsibilities. Agents really helpful to minimize the complexity of problem generated and scalability of agent help to making reliable decision and self organize the system.

3 Why? Number of reputed taxi company in colombo district – about 12 Number of taxies on colombo per day 01) budget taxies - more than 1100 02) car - more than 750 03) van - more than 250 Number of approximate hires per day in colombo – more than 8500 Approximate worth per day 12.5 million Witness of this situation – Last year (2015) Multi billions company of UBER, PickMe, HIRE1 and Vgo are established in Sri Lanka. All of these companies are technical companies but not traditional taxi companies.

4 Problems of traditional system
More than 6% of hire cancelation ratio per day. More than 34% of lost millage per vehicle. More than 45% of average waiting time per vehicle per hire. (27 min) More than 12% of average waiting time per customer per hire. (20 min) More than 47% of average time duration waiting on dispatch queue for dispatch per hire. (9 min)

5 Solution Steps – According to L. R.
Develop the vehicle tracking and monitoring system Develop the intelligent driver mobile app for communicate with drivers Develop the interactive client mobile app to getting client all requirement easily. 4) Multi agent base management system for manage the each and every expectable and un expectable situation base on prior knowledge.

6 Solution

7 Associated Technologies
NMEA, GPS data protocol for vehicle tracking and monitoring system Android Technology for driver and client mobile app development Google Cloud Messaging (GCM) architecture for server client communication Java Agent Development Environment (JADE) for managing complexity of this problem and re scheduling the exiting system without hassle.

8 Vehicle Tracking & Monitoring System

9 GPS data analysis for driver mobile app
Speed Graph of Taxi

10 GPS data analysis for driver mobile app
Ignition Graph of Taxi

11 State Analysis of Vehicle moving, Driver and Hire

12 State and Driver Messages
Moving State Driver State Hire State Message to Driver 1 None Are you free to getting hires? 2 3 Accept new job? 4 Are you at the place? 5 Did You start the hire? Any 6 Hire has been canceled! 7 Did you finish the hire? Are you going to waiting? 8 Did you stop waiting?

13 Intelligent Driver App

14 Interactive Client App

15 Agent Architecture

16 Automated Dispatching
Selecting taxi for dispatching Re arrange the taxis for economical dispatching Re-arrange the waiting taxis for new hires Apply the ride sharing to new bookings Apply the booking sharing to new bookings

17 Selecting Taxi for Dispatching - Agent

18 Selecting Taxi for Dispatching - Real
Taxi : Cab35209 EDF : Taxi : Cab43304 EDF : Taxi : Cab43814 EDF : 2.356 PL : # PT : :45:00 DL : # VT : Regular VM : None VIP : None Lan : English RD : None RS : 1 BS : 0 Taxi : Cab35181 EDF : 3.215 Taxi : Cab43815 EDF : 3.34 Taxi : Cab35176 EDF :

19 Selecting Taxi for Dispatching - Real
Taxi : Cab35209 EDF :1.3725 Taxi : Cab43304 EDF : Taxi : Cab43814 EDF : 2.356 PL : # PT : :45:00 DL : # VT : Regular VM : None VIP : None Lan : English RD : None RS : 1 BS : 0 Taxi : Cab35181 EDF : 3.215 Taxi : Cab43815 EDF : 3.34 Taxi – Cab35176 EDF –

20 Waiting Taxi arranging to New Hire - Real
ID : 1727 PL : # PT : :20:00 DL : # VT : None VM : None VIP : None Lan : None RD : None RS : None BS : None ID : 1684 PL : # PT : :10:00 DL : # DV : Cab43314 ASW : 1h 5min Taxi : Cab43314 ABI : 1684 CBI : 1727 EDF :

21 Waiting Taxi arranging to New Hire - Real
ID : 1727 PL : # PT : :20:00 DL : # DV : Cab43314 ASW : 0 min ID : 1684 PL : # PT : :10:00 DL : # DV : Cab43314 ASW : 1h 5min Taxi : Cab43314 ABI : 1727 PBI : 1684 CBI : None

22 Re-arranging Taxi - Agent

23 Re-arranging Taxi - Real
Taxi : Cab43315 ABI : 1745 CBI : None Taxi : Cab35181 ABI : 1734 CBI : None ERF1 : 3.456 ID : 1745 PL : # PT : :54:00 DL : # DV : Cab43315 ASW : 0 min ID : 1734 PL : # PT : :52:00 DL : # DV : Cab35181 ASW : 0 min

24 Re-arranging Taxi - Real
Taxi : Cab43315 ABI : 1734 CBI : None Taxi : Cab35181 ABI : 1745 CBI : None ID : 1734 PL : # PT : :52:00 DL : # DV : Cab35181 ASW : 0 min ID : 1745 PL : # PT : :54:00 DL : # DV : Cab43315 ASW : 0 min

25 Ride Sharing - Agent

26 Ride Sharing - Real AF : 3 ID : 1812 PL : 6.866503# 79.876399
PT : :02:00 DL : # VT : None VM : None VIP : None Lan : None RD : None RS : 1 BS : None DV : Cab35796 ID : 1814 PL : # PT : :03:00 DL : # VT : None VM : None VIP : None Lan : None RD : None RS : 1 BS : None Taxi : Cab35796 ABI : 1812 CBI : None

27 Ride Sharing - Real ID : 1812 PL : 6.866503# 79.876399
PT : :02:00 DL : # VT : None VM : None VIP : None Lan : None RD : None RS : 1 BS : None DV : Cab35796 ID : 1814 PL : # PT : :03:00 DL : # VT : None VM : None VIP : None Lan : None RD : None RS : 1 BS : None DV : Cab35796 Taxi : Cab35796 ABI : 1812, 1814 CBI : None

28 Booking Sharing - Agent

29 Booking Sharing - Real ID : 1871 ID : 1857 PL : 6.876952#79.864693
PT : :15:00 DL : # VT : None VM : None VIP : None Lan : None RD : None RS : None BS : 1 ID : 1857 PL : # PT : :14:00 DL : None VT : None VM : None VIP : None Lan : None RD : None RS : 0 BS : 0 DV : Cab35796 AF : 2 Taxi : Cab43814 ABI : None CBI : None Taxi : Cab35796 ABI : 1857 CBI : 1871

30 Booking Sharing - Real ID : 1857 PL : 6.877549#79.866410
PT : :14:00 DL : None VT : None VM : None VIP : None Lan : None RD : None RS : 0 BS : 0 DV : Cab35796 ID : 1871 PL : # PT : :15:00 DL : # VT : None VM : None VIP : None Lan : None RD : None RS : None BS : 1 DV : Cab35796 Taxi : Cab43814 ABI : 1871 CBI : None ID : 1871 PL : # PT : :15:00 DL : # VT : None VM : None VIP : None Lan : None RD : None RS : None BS : 1 DV : Cab43814 Taxi : Cab35796 ABI : 1857 CBI : None

31 Evaluation Booking cancellation ratio reduce to 8% to 0.6%
Average waiting time for dispatching reduce to 2 to 25 seconds Re arrangement of dispatched taxis increased to 0.9% to 8.1% Waiting vehicle arrange to new hires increased to 0.22% to 13.3% Ride sharing increase to 0.4% to 7.7% Booking sharing apply to 6.1% for all bookings.

32 Conclusion Referring historical data of client is powerful weapon arranging taxis Handling complexity and re scheduling the real time systems can manage with multi agent systems. Ride sharing within known client is good approach to resole the sharing limited resource. Booking sharing is good approach to reduce to client hail time on road and utilize the resources maximally.

33 Acknowledgement I am grateful to thank Radiant AC Cabs PVT LTD Company and their employees to providing their actual to us top to bottom. I am grateful to thank George Rzevski, (Professor Emeritus, Complexity Science and Design, The Open University, UK +44 (0) , for providing many tactic for managing above problem theoretically and practically.

34 Thank You & Questions ???


Download ppt "Multi Agent Approach for Time Based Vehicle Arrangement in Reputed Taxi Companies. 6037PS2013016 – D. A. P. Peiris Supervisors Mr. D.D.A. Gamini Mr. B."

Similar presentations


Ads by Google