ENCE 688R Projects, Spring Semester, 2017

[ Project 1 ]: Simplified Hotel Reservation System
[ Project 2 ]: TripPlanner Data Model
[ Project 3 ]: Reinforcement Learning for the game Flappy Bird
[ Project 4 ]: Simulation of Airplane Trajectories
[ Project 5 ]: Natural Language Processing and Feature Extraction of Sentences
[ Project 6 ]: Multi-Threaded Transfer Money System
[ Project 7 ]: Long Distance Travel Model
[ Project 8 ]: ToDo List App
[ Project 9 ]: Prognostics and Health Management Software


Title: Simplified Hotel Reservation System
Authors: Hua Luo

Abstract: In the past few decades travelling has become more and more common to everyone. People travel for business, to visit friends, or to explore other parts of the world as well as get to know themselves.

An essential and important part of the trip is booking hotels. Companies such as Airbnb, Booking.com, and TripAdvisor have made a lot of effort to make this process more convenient for everyone.

In this project, we will design and simulate a simple hotel-booking system using Python. The program will read hotel data from an input file and initialize the hotel system. With this system, users can check the availability of all hotels for a given check-in and check-out date. They can also make reservations at a specific hotel.



Title: TripPlanner Data Model
Author: Ruiche Liu

Abstract: Nowadays, planning a trip can be exhausting. People spend much time to schedule their trips and get all the activities organized. This program can help people to make trip plan easier. Once people get their trip plan in an xml file, this program can read xml file and sort all activities in a time series. Then it can print all activities in a table form and a map view. For the further calculation, the program can provide a summary about the plan, for example, travel distance, activity duration and total cost.

Figure 1. System Architecture for a Trip Planner Application.

Figure 1 shows the system architecture and the process of data transmission.



Title: Reinforcement Learning for the game Flappy Bird
Author: Tingzhen Du

Abstract: Reinforcement learning has been widely exposed to the public since Alpha Go. The performance of Alpha Go is remarkable because the game and goals were even challenging for people. Since the paper ``Playing Atari with Deep Reinforcement Learning'' was published, reinforcement learning has been widely used on different games without changing the core part of the network. The powerful point of reinforcement learning is it could be applied to tasks where there is no single correct way to solve the problem.

In this project, reinforcement learning application on a game called flappy bird will be discussed.

Key words: Deep Q Network, Reinforcement Learning, Tensorflow, Amazon AWS, Convolutional Neural Network, OpenCV.



Title: Simulation of Airplane Trajectorie
Author: Kenan Unal

Abstract: Goal of the project will be to create a java program that can simulate the travel of several airplanes flying over the United states when given a .xml file containing properties of the airplanes. e.g., the starting and landing airports, the airplane speed, the expected travel time, and the height it will be flying at.



Title: Natural Language Processing and Feature Extraction of Sentences
Author: Linyu Shi

Abstract: My project is to build a simple natural language processing system, based on Apache OpenNLP library. The input for my system would be several sentences, and my system would tokenize the sentence and identify some special field of the whole sentence, such as the physical unit and day of week. In addition, my system would classify the sentence according to these specific fields.

To implement the system, I will pre-define some dictionary like a wordnet for these specific field. For the classification, the system may answer the query like ``Which sentence is talking about Maryland?'' or classify the input sentences into several groups respected to the user query like speed.



Title: Multi-Threaded Transfer Money System
Author: Xi Wu

Abstract: In this project, I will design a bank transfer money system. It allows multi-thread transfer money. To improve the efficiency during multi-thread transferring, different types of lock will be designed and used to achieve it. Moreover, deadlock detection will be designed to detect if there is any deadlock in bank system. In addition, recovery system will be considered to avoid system crush during transferring money. A log file will be kept for each transferring to do the recovery. Test cases will be designed to test the bank system.



Title: Long Distance Trip Model
Author: Arash Asadabadi

Abstract: Long Distance trip model is an activity based nationwide model for predication the long distance trips for each person in USA based on his/her socio-economic inputs (The model produces trip counts for each quarter of year and also identifies the travel mode choices by each person). I am working on two scenarios in Maryland area, e.g. How constructing the new casino in DC will affect people travel mode choice in this area. This is doing by implementing some changes in both input land use files and also the mode choice module in the model. The model is fully object oriented making it easier to implement changes in different modules of the model.



Title: ToDo List App
Author: Mingxi He

Abstract: In this project, I will design a TODO app that is able to edit and display the user's todo list, including basic information, date, time, and progress (whether completed).

Functionalities like auto-ordering the TODO list, updating the completion status, changing the date or time and adding or deleting one todo schedule will be included.

I will use ListView that applies the adapter to display the view. Maybe I will examine some changes in layout (i.e., appearance) and add an alarm functionality for improvement.


Title: Prognostics and Health Management Software
Author: Varun Khemani

Abstract: The intent of the project is to develop parts of a software to help users carry out the following steps:

  1. Data Pre-Processing
  2. Feature Discovery
  3. Anomaly Detection
  4. Diagnostics Classification
  5. Prognostics/Modeling

It is not intended to develop the entire software as it will involve multiple algorithms in each step. For the purposes of this project, only algorithms related to graphs and/or networks will be developed. This may include all/some of the following algorithms:

  1. Regression Trees
  2. Random Forests
  3. Self-Organizing Maps

The algorithms will be coded in Java or Jython using Java Collections. The software GUI will be developed using JavaFx.

Developed in April 2017 by Mark Austin
Copyright © 2017, Department of Civil and Environmental Engineering, University of Maryland