21
AN EFFICIENT PREFERENCE-BASED PERSONAL TASK SCHEDULER Gabriel John P. Gagno November 24, 2014 Institute of Computer Science Francisco O. Santos Hall University of the Philippines Los Baños

Special Problem Presentation

Embed Size (px)

DESCRIPTION

Presentation of "An Efficient Preference-Based Personal Task Scheduler", a Special Problem by Gabriel John P. Gagno of the University of the Philippines Los Banos

Citation preview

Page 1: Special Problem Presentation

AN EFFICIENT PREFERENCE-BASED PERSONAL TASK SCHEDULERGabriel John P. GagnoNovember 24, 2014Institute of Computer ScienceFrancisco O. Santos HallUniversity of the Philippines Los Baños

Page 2: Special Problem Presentation

INTRODUCTION

Page 3: Special Problem Presentation

BACKGROUND

Scheduling is key to beating real-world deadlines. Everyone does some way of scheduling as a way to plan their

activities in a day. Scheduling allows people to plan ahead their activities with

efficiency and ease.

Page 4: Special Problem Presentation

BACKGROUND

The main aim of scheduling is to find a way to minimize submissions after the deadline (avoid being late).

Scheduling also aims to minimize idle time and keep the person productive at all times.

Page 5: Special Problem Presentation

BACKGROUND

Scheduling itself takes time. Currently, scheduling personal tasks are done manually. People spend time creating plans/schedules first before doing

the tasks.

Page 6: Special Problem Presentation

OBJECTIVES

The study aims to devise an algorithm that organizes tasks according to the user’s preference and organize them into a timetable.

Find an algorithm that arranges tasks into a list satisfying the user’s preferences

Dynamically assign tasks into a timetable. Build an application that implements the devised algorithm. Measure user satisfaction of the application (based on the

generated schedules).

Page 7: Special Problem Presentation

SIGNIFICANCE

This study aims to benefit students and working people. Saves time Allows them to start working instead of planning first Less energy spent on planning

Page 8: Special Problem Presentation

PROPOSED SOLUTION

Page 9: Special Problem Presentation

A GENERAL NOTATION FOR SCHEDULING

The scheduling problem in this paper’s context can be defined as with the triplet

is the machine environment indicates a release date indicates that preemption is allowed The objective is to minimize the tardiness in all of the schedule

Page 10: Special Problem Presentation

JOBS, TASKS, AND HABITS

Jobs are the inputs that a single machine (i.e. person) has to do. It has the following properties:

Processing time Release date

The application divides the general concept of jobs in scheduling theory into two: tasks and habits.

Tasks are jobs that have a deadline, a defined number of operations, and a duration. These tasks are the ones that could be scheduled.

Habits are jobs that do not have the same properties as tasks, but they have an assigned frequency. These jobs represent real-life jobs that are repeatable over time.

Page 11: Special Problem Presentation

JOB WEIGHT

Each task is assigned its own weight. Weights are to facilitate the ranking of the tasks in the priority

queue. The weight formula is denoted by

indicates the time constant is the due date is the current date is the constant for processing time is the processing time

Page 12: Special Problem Presentation

PRELIMINARY RESULTS

Page 13: Special Problem Presentation

THE APPLICATION

Reverie is a Java-based application that generates schedules based on these user preferences:

Deadline Completion time

Users can input tasks and habits Users can set their preferences Could run in the background

Page 14: Special Problem Presentation

INITIAL SCHEDULE

We provide a pre-built schedule with the following task already in store:

Job Title: CMSC150 Job Note: Optimize Job Deadline: 11/28/2014 4:00 PM Number of Operations: 1 Duration of each operation: 3

This job is initially assigned the following dates in the schedule: Start time: 11/24/2014 5:10 AM End Time: 11/24/2014 8:10 AM

Page 15: Special Problem Presentation

INITIAL SCHEDULE

Time Sunday Monday Tuesday Wednesday Thursday Friday Saturday

5-66-77-88-99-1010-1111-1212-11-2

Do CMSC150

Page 16: Special Problem Presentation

INITIAL SCHEDULE

We provide three sample inputs, shown in the next slide. All three inputs have one operation only.

Page 17: Special Problem Presentation

INITIAL SCHEDULE

Input 1 Job Title: do SP Job Deadline:

11/24/2014 5:00 PM

Operation Duration: 3

Input 2 Job Title: laundry Job Deadline:

11/24/2014 5:00 PM

Operation Duration: 1

Input 3: Job Title: do 137 Job Deadline:

11/24/2014 5:00 PM

Operation Duration: 2

Page 18: Special Problem Presentation

TEST RUN

Running them in the application, we arrive at the following schedule.

jobId, job start, job end

Page 19: Special Problem Presentation

TEST RUN

Time Sunday Monday Tuesday Wednesday

Thursday

Friday Saturday

5-66-77-88-99-1010-1111-1212-11-2

laundry

Do 137

Do SP

Do CMSC 150

Page 20: Special Problem Presentation

PROTOTYPE STATUS

Weight model is being continuously improved (and proven). Algorithm is being continuously improved to eventually allow

multi-operational jobs (i.e. allowing preemption/staggering). Graphical, more user-friendly interface is not yet available.

Page 21: Special Problem Presentation

END