Date of Award

9-18-2018

Publication Type

Master Thesis

Degree Name

M.Sc.

Department

Computer Science

Keywords

Cluster Hire, Communication Cost, Experts, Network, Profit

Supervisor

Chen, Dr. Jessica

Supervisor

Kargar, Dr. Mehdi

Rights

info:eu-repo/semantics/openAccess

Abstract

Finding a group of experts is a natural way to perform a collection of tasks that need a set of diversified skills. This can be done by assigning skills to different experts with complementary expertise. This allows organizations and big institutes to efficiently hire a group of experts with different skill sets to deliver a series of required tasks to finish a set of projects. We are given a collection of projects, in which each of them needs a set of required skills. Performing each project brings a profit to the organization. We are also given a set of experts, each of them is equipped with a set of skills. To hire an expert, the organization should provide her with monetary cost (i.e., salary). Furthermore, we are given a certain amount of budget to hire experts. The goal is to hire a group of experts within the given budget to perform a subset of projects that maximize the total profit. This problem is called Cluster Hire and was introduced recently. We extend this problem by making the realistic assumption that there exists an underlying network among experts. This network is built based on past collaboration among experts. If two experts have past collaboration, they form a more collaborative and efficient team in the future. In addition to maximizing the total profit, we are also interested to find the most collaborative group of experts by minimizing the communication cost between them. We propose two greedy algorithms with different strategies to solve this problem. Extensive experiments on a real dataset show our proposed algorithms can find a group of experts that cover projects with high profit while experts can communicate with each other efficiently.

Download

Share

COinS