Method and system for simulating users in the context of a parking lot based on the automatic learning of a user choice decision function from historical data considering multiple user behavior profiles

• US 9,576,250 B2
• Filed: 02/24/2015
• Issued: 02/21/2017
• Est. Priority Date: 02/24/2015
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: de Souza, Cesar Roberto, Roulland, Frederic, Ciriza, Victor
• Primary Examiner(s): St Cyr, Daniel
• Application Number:
  US14/630,227
• Publication Number:
  US 20160247326A1
• Time in Patent Office: 728 Days
• Field of Search: 235/376, 235/380, 235/383, 235/384, 235/385
• US Class Current: 1/1
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/067 Enterprise or organisation modelling
• CPC Class Codes: G06Q 30/0284 Time or distance, e.g. usage of parking meters or taximeters
• CPC Class Codes: H04L 67/306 User profiles

Abstract

Methods and systems for modeling user arrival and choice in the context of off-street parking solutions. A first component models the arrival and duration of stay of users as a function of time, taking into account different user profiles (or “clusters”), captured by a latent variable. A second component provides a ranking function (for each user cluster), wherein the input features describing the “choice” constitute status variables associated different car park(s), and the output constitutes a preferred car park and a pricing scheme. The system simulates different user behaviors by assuming some standard groups of users will behave similarly. Groups of users or user profiles are learned automatically. The profiles are then employed as a key element for automatically learning a decision function of parking users, and automatically learning one decision function per profile.

Claims

1. A system for use in managing an off-street parking lot, comprising:

at least one user interface; and

at least one module for micro-simulation of parking users of at least one parking area, wherein said micro-simulation of said parking users of said at least one parking area is used to predict results of price rate changes, infrastructure changes, and service changes with respect to said at least one parking area, wherein said micro-simulation comprises a simulating parking environment that is graphically visualized and accessible through said at least one user interface, wherein said at least one user interface is accessible to edit price grids, invoke and test simulations, and produce outcome reports comparing simulated world data against real-world data.

2. The system of claim 1 wherein said at least one module for said micro-simulation of parking users comprises:

a database module that stores and maintains historical data for use in said micro-simulation of said parking users of said at least one parking area;

a user demand module that communicates electronically with said database module, and which estimates when a new parking user among said parking users of said at least one parking area will arrive at said parking lot, wherein said historical data is used to build said user demand module, and wherein said user demand module implements a generative approach for simplifying a task of generating entries with respect to an initialized demand model employed in said micro-simulation of said parking users of said at least one parking area;

a user choice module that estimates a choice in parking that may be made by a parking user among said parking users of said at least one parking area;

a price engine module that communicates electronically with said user choice module and computes how much said parking user will be required to pay if said parking user opts for said choice in parking;

a virtual parking module that provides a virtual parking lot; and

a simulation module that communicates electronically with said user demand module and said database, and which builds said virtual parking lot based on said historical data and generated by said user demand module, said user choice module, and said price engine module.

3. The system of claim 2 wherein said at least one module:

optionally divides said parking users into distinct groups that behave similarly so as to provide enhanced predictions regarding behaviors associated with said parking users with respect to said at least one parking area;

wherein each user profile associated with each user among said parking users is assigned a different choice model from among a plurality of choice models learned from historical data and other factors via a supervised learning module.

4. The system of claim 3 wherein said user demand module is based on a non-parametric decomposition of arrival and duration times for a configurable time period that encompasses any fixed number of weeks, months, or years.

5. The system of claim 4 wherein probability distributions describing different features of each user profile associated with said parking users are combined to form a single probability distribution in which a first component comprises a cluster probability.

6. The system of claim 4 wherein said data is leveraged in a feature extraction process to teach different user choice models.

7. The system of claim 2 wherein said user choice model comprises a machine learning algorithm utilizing derived measures estimated as a part of a learning process after learning databases have been created.

8. The system of claim 7 wherein said at least one module additionally employs subjective data available to at least one parking user among said parking users of said at least one parking area at a time a choice is made by said at least one parking user, wherein said subjective data is employed by said at least one module for said micro-simulation of said parking users.

9. The system of claim 2 wherein said micro-simulation comprises a discrete event micro-simulation and said at least one module:

creates learning data;

for said micro-simulation of said parking users of said at least one parking area at each event of said discrete event micro-simulation;

extracts a positive feature vector for each choice made by at least one parking user among said parking users of said at least one parking area; and

generates a negative feature vector for each choice not made by said at least one parking user among said parking users of said at least one parking area.

10. The system of claim 9 wherein said at least one module extracts a plurality of feature vectors for use in said micro-simulation by running said micro-simulation with real data and registering a current status of said at least one parking area at each instant of said discrete event micro-simulation.

13. The system of claim 2 wherein said instructions are further configured for:

optionally dividing said parking users into distinct groups that behave similarly so as to provide enhanced predictions regarding behaviors associated with said parking users of said at least one parking area with respect to said at least one parking area; and

wherein each user profile among said user profiles associated with said parking users is assigned a different choice model from among a plurality of choice models learned from said historical data via a supervised learning module that considers different pricing changes in said historical data.

14. The system of claim 13 wherein:

said plurality of choice models are learned from an augmented dataset that has been reconstructed from a database of historical records of said at least one parking area; and

subjective data available to at least one parking user among said parking users at a time a choice is made by said at least one parking user is employed for learning choice models among said plurality of choice models employed in said micro-simulation of said parking users of said at least one parking area.

15. The system of claim 13 wherein said instructions are further configured for extracting the most important features guiding a decision of at least one user among said parking users of said at least one parking area in each of said user profiles utilizing an L1-regularized linear model comprising at least one of 1-regularized logistic regression or an L1-support vector machine.

16. The system of claim 13 wherein said instructions are further configured for:

estimating data indicative of inter-zone price elasticity information for different user profiles among said user profiles; and

incorporating said data into a parking visualization tool.

11. A system for use in managing an off-street parking lot, comprising:

at least one processor; and

a computer-usable medium embodying computer program code, said computer-usable medium capable of communicating with said at least one processor, said computer program code comprising instructions executable by said at least one processor and configured for;

providing at least one user interface; and

generating a micro-simulation of parking users of at least one parking area, wherein said micro-simulation of said parking users of said at least one parking area is used to predict results or price rate changes, infrastructure changes, and service changes with respect to said at least one parking area, said micro-simulation comprises a discrete event micro-simulation, wherein said micro-simulation comprises a simulating parking environment that is graphically visualized and accessible through said at least one user interface, wherein said at least one user interface is accessible to edit price grids, invoke and test simulations, and produce outcome reports comparing simulated world data against real-world data.

12. The system of claim 11 further comprising:

a database module that stores and maintains historical data for use in said micro-simulation of said parking users;

a user demand module that communicates electronically with said database module, and which estimates when a new parking user among said parking users of said at least one parking area will arrive at said parking lot, wherein said historical data is used to build said user demand module, and wherein said user demand module implements a generative approach for simplifying a task of generating entries with respect to an initialized demand model employed in said micro-simulation of said parking users;

a user choice module that estimates a choice in parking that may be made by a parking user among said parking users of said at least one parking area;

a price engine module that communicates electronically with said user choice module and computes how much said parking user of said at least one parking area will be required to pay if said parking user opts for said choice in parking;

a virtual parking module that provides a virtual parking lot; and

a simulation module that communicates electronically with said user demand module and said database, and which builds said virtual parking lot based on said historical data and generated by said user demand module, said user choice module, and said price engine module.

17. The system of claim 11 wherein said instructions are further configured for:

creating learning data for said micro-simulation of said parking users at each event of said discrete event micro-simulation;

extracting a positive feature vector for each choice made by at least one parking user among said parking users; and

generating a negative feature vector for each choice not made by said at least one parking user among said parking users.

18. A method for managing an off-street parking lot, comprising:

providing at least one user interface;

generating a micro-simulation of parking users of at least one parking area; and

employing said micro-simulation of said parking users of said at least one parking area to predict results of price rate changes, infrastructure changes, and service changes with respect to said at least one parking area, wherein said micro-simulation comprises a simulating parking environment that is graphically visualized and accessible through said at least one user interface, wherein said at least one user interface is accessible to edit price grids, invoke and test simulations, and produce outcome reports comparing simulated world data against real-world data.

19. The method of claim 18 further comprising:

creating learning data for said micro-simulation of said parking users of said at least one parking area at each event of said micro-simulation;

extracting a positive feature vector for each choice made by at least one parking user among said parking users of said at least one parking area; and

generating a negative feature vector for each choice not made by said at least one parking user among said parking users of said at least one parking area.

20. The method of claim 18 further comprising creating said micro-simulation of parking users utilizing:

a database module that stores and maintains historical data for use in said micro-simulation of said parking users;

a user demand module that communicates electronically with said database module, and which estimates when a new parking user among said parking users will arrive at said parking lot, wherein said historical data is used to build said user demand module, and wherein said user demand module implements a generative approach for simplifying a task of generating entries with respect to an initialized demand model employed in said micro-simulation of said parking users;

a user choice module that estimates a choice in parking that may be made by a parking user among said parking users;

a price engine module that communicates electronically with said user choice module and computes how much said parking user will be required to pay if said parking user opts for said choice in parking;

a virtual parking module that provides a virtual parking lot; and

a simulation module that communicates electronically with said user demand module and said database, and which builds said virtual parking lot based on said historical data and generated by said user demand module, said user choice module, and said price engine module.