Date of Award

May 2021

Degree Type

Thesis

Degree Name

Master of Science

Department

Computer Science

First Advisor

Amol Mali

Committee Members

Rohit Kate, Mohammad Rahman

Abstract

In the current era, Diversity, Equality, and Inclusion (DEI) are often not sufficiently addressed due to bias against certain people or unjust stereotypes or simply an inadequate understanding of the value of DEI. Not addressing DEI sufficiently leads to multiple problems including lawsuits, costly settlements, departure of valuable employees, reduced employee productivity, shortage of qualified workforce, and unjust hiring, compensation, and work-distribution practices. Initiatives to address DEI often fail or risk being ineffective. In this thesis, Food Delivery and Software Company domains are introduced by modeling assignment of drivers to customers who ordered food, and by modeling hiring and project assignment in software industry, using Planning Domain Definition Language (PDDL). This thesis is the first research work to address DEI using fully automated symbolic planning. The experimental results obtained on multiple instances on the 2 domains show that it is possible to express DEI-related objectives using PDDL, and get plans conforming to these objectives. Discrimination is addressed indirectly by nine of the eleven versions of the two domains by enforcing DEI-related objectives. This work can be extended or adapted for use in other domains involving planning or scheduling, to address DEI and discrimination in multiple ways at multiple levels, and the expressive power of PDDL and efficiency of domain-independent as well as domain-specific planners can be exploited in the process.

Available for download on Wednesday, August 30, 2023

Share

COinS