1. Customized Simulation Modeling Using PARAMICS Application Programming Interface Henry Liu and Lianyu Chu

2. Outline Paramics System Overview PARAMICS API Development Adaptive Ramp Metering Study Demo

3. Paramics System Overview

4. PARAMICS (PARAllel MICroscopic Simulation) is a suite of software tools for microscopic traffic simulation, and represents a radical new approach to the understanding, representation and analysis of road traffic.

5. Components... PARAMICS Modeler PARAMICS Processor PARAMICS Programmer PARAMICS Analyzer

6. Modeler Core Simulation and Visualization Tool Provide Fundamental Operations Aspects of Transportation Network 1. freeway networks  Right-hand and left-hand drive capabilities 3. Advanced signal control 4. Roundabouts 5.  Public transportation 6.  Car Parking 7. Incidents 8. Truck-lanes, high occupancy vehicle lanes

7. Processor Tool for batch mode, especially multiple simulations run Reduce simulation time greatly, without visualizing network and vehicles.

8. Analyz er Analysis tool for displaying the output and report on statistical data. (on-screen display, text file, spreadsheet, databases…) A range of statistics such as: vehicle paths, traffic flow volumes by link and turn, queue length, intersection delay, etc.

9. Programme r Customized Paramics Using Application Programming Interfaces (API). Override the internal models such as car following, lane changing, and route choices, etc. Interface complementary modules such as signal optimization, adaptive ramp metering, and incident management, etc. Eventually, Paramics Simulation Shell.

10. Advantages... Scalability: No limits on network size Intuitive graphical network editor Simultaneous visualization and simulation Comprehensive statistical output Integrated urban and freeway modeling

11. Features... Detailed network coding, including horizontal curves, superelevation, other complex geometrics Extendable: easy use cut-and-paste network editing. WYSIWYG (what you see is what you get).

13. Features... Microscopic representation of both vehicle and driver. Configurable behavior and attributes of driver and vehicle

15. Features... Comprehensive visual display Effective, easy to use GUI Highly detailed and extensive network definition Complex junction modeling: pre-timed signal, user-definable logic for actuated control, interface for type 170 controller

17. Features... Extensive Output analyzer Comprehensive simulation output at aggregate and disaggregate levels Statistical analysis of user-specified parameters and MOEs

18. Features… Traffic assignment: all-or-nothing, stochastic, dynamic feedback Intelligent routing functionality Application Programming Interface (API) Integrated simulation of ITS elements

19. Features… Transit Modeling: user-definable routes, timetables and frequency, bus stop location, capacity, arrival rate of passengers, bus priority signals, etc. Incident modeling: user-definable incident type, location, time, duration, severity.

20. Additional Features … Car Park modeling Demand modeling Pollution modeling Pedestrian and cyclist modeling

21. is... Validated: used in UK, Germany, France, Japan, Singapore, Taiwan, Argentina, Canada, and US. Portable: ready for different unit systems (UK, US, and metric) and available on a variety of computing platforms (Windows95, 98, NT, Sun, SGI, and HP).

22. http://www.paramics-online.com/ Additional Information …

23. Customized Paramics Using API

24. User Developer Output Interface Input Interface GUI Tools Professional Community Oversight Core Model API Introduction

25. Introduction (Contd.) API provides users with a functional interface Command- based With GUI With API Data Interface Functional Interface Simulation Program

26. Introduction (Contd.) function calls: vehicle related.. link related.. and others user-defined programs Main simulation loop Plugins API data Other applications /APIs Role of a typical API functions

27. Introduction (Contd.)  Customization pushing the limits More on the API…  Plug-and-play environment reusable and generic plugins  API: the “soft key” to the black-box

28. Why Customize? Incident Detection Intelligent Parking Travel Time Prediction Signal Control Systems Transit Priority Electronic Road Pricing Road Maintenance Scheduling & Monitoring Bus Scheduling Assistance TESTBED

29. Why Customize? (Contd.) Network Building Performance Measurement Additional Functionality: ITS Elements Basic Functionality: signals etc Customize

30. PARAMICS API Simulation Loop Overload Functions Override Functions Callback Functions Built-in Functions User Functions

31. PARAMICS API (Contd.) Access via API At every timestep (or at intervals) When an event occurs in simulation Event triggered by user

32. PARAMICS API Development A Hierarchical Approach Provided API Library ATMIS Modules Developed API Library Advanced Algorithms Data Handling Routing Ramp Signal CORBA Databases Adaptive Signal Control Adaptive Ramp Metering Dynamic Network Loading Demand XML

33. Developed Basic API Library Path-based Routing (Para-Dyn) Paramics-CORBA Communication Actuated Signal Controller Time-based Ramp Metering Paramics-MySQL Communication Loop Aggregator Performance Measurement

34. Modules Developed Actuated Signal Control Plugin Inputs: Signal Timing Plan, including phase sequence, initial green, maximum green, unit extension time and system recall phase, etc. Detectors need to be specified and associated with movements to be activated. Standard During-ring Logic Actuated Signal Coordination Interface with some other signal optimization packages such as Transyt7F and SYNCHRO, etc.

35. Modules Developed Ramp Metering Basic Time-based Module: Input: time-of-day ramp control plan such as 6-9 AM, cycle length 5 sec. Logic: n-cars-per-green Advanced Modules: Demand-capacity strategy Percent-occupancy strategy ALINEA BOTTLENECK ZONE

36. Utility Plugins Developed Paramics-MySQL Communication Connecting PARAMICS simulation environment with MYSQL database Includes a set of simple C routines programmed in MYSQL API functions. The MYSQL database can be used in the following two folds: API users can store the simulation outputs to database; During a simulation process, MYSQL database can be used for storing intermediate simulation results, such as aggregated loop data, which can be queried by other external API modules at any time.

37. Utility Plugins Developed Loop Aggregator Input: time interval, smooth factor, detector name Output: MYSQL database or ASCII file volume, percent occupancy, speed, flow, headway

38. Performance Measurement Plugin Utility Plugins Developed To customize performance measurement for run-time interfacing with other tools such as data mining and signal optimization. MOE: vehicle count, travel time, stopped time, vehicle-spent time in a specific speed range, turn counts from intersections, cycle time, individual phase time etc. Data collected at a detector, node, link, corridor, OD pair or network levels, at specified time intervals, for specific type of vehicles where applicable. Output can be in the form of a spreadsheet, text file or on-screen reporting.

39. Wrap up 1. While GUI helps in building a basic simulation network, API helps in customization of various functional aspects of simulation modeling. 2. Plugins provide users with more freedom to interrupt and control simulation processes and hence facilitates overcoming some of the challenges faced in modeling traffic scenarios of the ITS era.

40. Publications 1.Liu, X., Chu, L., and Recker, W., “Paramics API Design Document for Actuated Signal, Signal Coordination and Ramp Control”, California PATH Working Paper, UCB-ITS- PWP-2001-11, University of California at Berkeley, 2001. 2.Chu, L., Liu, X., Recker, W., and Zhang, H. M., “Development of A Simulation Laboratory for Evaluating Ramp Metering Algorithms”, Accepted for the presentation at TRB 2002.

41. Case Study: Adaptive Ramp Metering Evaluation

42. Categories of Ramp Metering Algorithms Fixed-time Local traffic responsive, and Coordinated traffic responsive

43. Ramp-metering algorithms implemented in PARAMICS API Fixed-time Demand-capacity strategy Percent-occupancy strategy Alinea, local feedback control policy BOTTLENECK ZONE

44. Framework of simulation laboratory

45. PARAMICS Calibration Network Geometry Driver behavior factors Vehicle characteristics, the proportion of vehicle types OD demands Compared Loop counts of simulation with those in the real world Comparison of Speed Contour Maps Speed-Flow Curve

46. EVALUATION STUDY Study site

47. Measures of Effectiveness 1.Total system travel time. 2.Total traffic throughput of the mainline freeway 3.Mainline average travel time. 4.Average on-ramp waiting time and average queue length in number of vehicles on the entrance ramp. 5.OD travel time

48. Overall system performance

49. Average waiting time (AM)

50. Average waiting time (PM)

51. Mainline Travel time (AM)

52. OD Travel time (AM)

53. Evaluation results The two coordinated ramp-metering algorithms, i.e., BOTTLENECK and ZONE, are more efficient than the current fixed-time control and ALINEA under both morning scenario and afternoon scenario. The ZONE algorithm performs best among these three adaptive ramp-metering algorithms. ALINEA performs worse than the fixed-time control under the morning scenario, but shows a better performance under the afternoon scenario.

54. Contact Information: hliu@translab.its.uci.edu lchu@translab.its.uci.edu Thank You!