Theses and Dissertations from UMD
Permanent URI for this communityhttp://hdl.handle.net/1903/2
New submissions to the thesis/dissertation collections are added automatically as they are received from the Graduate School. Currently, the Graduate School deposits all theses and dissertations from a given semester after the official graduation date. This means that there may be up to a 4 month delay in the appearance of a give thesis/dissertation in DRUM
More information is available at Theses and Dissertations at University of Maryland Libraries.
Browse
3 results
Search Results
Item LARGE SCALE AGENT-BASED MODELING: SIMULATING TWITTER USERS(2016) Ariyaratne, Arjuna; Rand, William; Systems Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This thesis details an attempt to conduct a large-scale Agent-based modeling simulation where simulating Twitter users are used as an example. In this thesis, Computational Mechanics is used for developing rules that govern each Agent. This thesis also details the development of an entirely new simulation software capable of simulating a large number of Agents by taking advantage of the parallelism offered by latest computing platforms. Development details of this simulation software, named elixrABM, is available from the concept phase to the testing phase of the simulator.Item A Complexity-Based Approach to Intra-Organizational Team Selection(2013) Hsu, Shu-Chien; Cui, Qingbin; Skibniewski, Miroslaw; Civil Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Early studies recognized the significance of team's work capacity and suggested the selection of team members based on individual skills and performance in alignment with task characteristics. The equitable team selection method, for example, assigns people to different tasks with even skill distributions for the best overall performance. Recent advancement in organization science also identifies the importance of contextual skills. However, work teams are complex adaptive systems with interdependence between workers and social environment, and exhibit surprising, nonlinear behavior. Optimizing individual stages without taking organizational complexity into account is unlikely to yield a high performing new combination of teams. The objectives of this study can be stated as: a) Utilizing complex system theory to better understand the processes of team selection including forming teams with considering worker's interdependence and replacing the unsuitable members through a time frame; b) Comparing different team selection methods, including random selection, equity method, using knowledge of interdependence in different economic conditions through simulation; c) Comparing different policies of replacing members of teams. This study utilizes a computational model to understand the complexity of project team selection and to examine how diversity of capability and interdependence between workers to effect team performance in different economic conditions. The NK model, a widely used theory for complex systems is utilized here to illustrate the worker's interdependence and fed into an Agent-Based Model. This study uses a small design firm as a case implementation to examine the performance of a variety of team selection approaches and replacement policies. Project data, task assignment, and individual and team performance information were collected for the period of 2009-2011. The simulation results show that while the equity selection method can increase the diversity of capabilities of teams, the net performance is often worse than optimizing worker interdependencies. This study suggests that managers should protect their higher-performing workers with minimal interdependence disruption when they considered team selection. Thus taking the advantages and disadvantages of all three policies into account, transferring low contributors or least supported members are recommended to be enacted before hiring new workers to avoid this last policy's especially large additional costs.Item The Effect of Economic Condition on Civil Unrest: New Insights from Agent Based Modeling(2013) Harry, Charles Thomas; Steinbruner, John; Public Policy; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)On the morning of December 17th 2010, a Tunisian fruit vendor named Mohamed Bouazizi took a can of gasoline and lit himself on fire in front of the local governor's office in protest of having his cart and its inventory confiscated by a corrupt police force. What began as act of individual protest against the Tunisian state evolved into mass demonstrations across the Arab world significantly altering the political landscape. While researchers have explored links between economic determinants and the outbreak of civil conflict, use of large datasets and econometric analysis alone has been unable to draw the indelible connections scholars had hoped. The use of country wide data are seen as insufficient in capturing localized dynamics of civil violence calling into question the applicability of current conceptual frameworks. Case study and data collection at sufficient granularity are now necessary steps in the further exploration of this topic. To help develop our understanding of how localized conflict emerges this dissertation steps away from the use of traditional case study and econometric analysis to develop two agent based models of protest that allow exploration to changes in the average level of utility, its distribution, and growth on the onset, size, and frequency of protest. This dissertation finds that the level and distribution of utility affects the onset, evolution, magnitude, and frequency of civil instability. Further, empirical methods explored to date fail to account for significant micro dynamic behavior that influences the emergence of mass protest. This paper extends earlier modeling work on civil protest and discusses findings on how utility levels in a system of agents affect the magnitude, frequency, onset, and structure of civil conflict. The most surprising finding in this paper is the bifurcated relationship between utility distribution and the emergence of civil conflict. This specific result provides a plausible explanation for why empirical analysis has thus far been unable to correlate income inequality and civil violence.