Arpi Mardirossian

In the ISE 530: Introduction to Operations Research course that I teach at USC, the students recently completed class projects. They all did a fantastic job! Take a look at their presentations:

Team Masters of Blackjack discovered a method for playing blackjack that would dramatically improve winnings when playing blackjack. The method consists of following existing guidelines or cheat sheets, always starting with a minimum required bet, and doubling the next bet after a loss. Click here to view their presentation.

Team Innovative Slackers developed an application to help with making the decision of buying verses leasing a car. Their approach incorporates not only financial considerations, but also vehicle and lifestyle preferences. Click here to view their presentation.

Team Four Constraints did a case study of a local Coffee Bean to determine whether the number of staff members employed was optimal. They discovered that Coffee Bean was over staffing and they could, in certain shifts, reduce their workforce in half! Click here to view their presentation.

Team Crime Alert studied the archives of crime reports for USC and the surrounding areas in order to predict the hot spots for future crime. They discovered three distinct areas with a high crime rate and determined which one of these areas is most likely to have a high rate of crime. Click here to view their presentation.

Team Lords of OR studied the movie recommendation algorithm used by Netflix and proposed a new, improved algorithm that incorporates collaborative filtering with a content-based approach. Click here to view their presentation.

Team Fantastic Four designed a new and improved shuttle bus system for the USC campus. They determined the optimal number and locations of shuttle stops, routes and schedule. Click here to view their presentation.

Team Workers Union optimized the process of worker wage determination. They clustered workers into groups and dynamically offered each group different wages based worker group desirability and on demand for work. Click here to view their presentation.

I am very proud to announce the launch of Tagasauris. We have been working hard to develop innovative solutions for unlocking the value of multimedia collections by improving find-ability through quick, cost-effective and accurate descriptive labels. Tagasauris is a media tagging tool that’s powered by the crowd. We provide superior quality tags at lightening speed for a fraction of the cost of more traditional methods. We also connect these tags to well defined concepts and metadata with unique URLs. This means that everybody uses the same names for tags from the world’s largest collection of knowledge.

I have been working with Tocarte on the development of a high performance annotation solution that makes images more discoverable and connected. The system also helps optimize both productivity and accuracy of crowd-sourced workers. It goes beyond simple free-text tags and associates strong IDs to unique, well-defined concepts and metadata.