Matching Items (53)
Description
The Cognitive Decision Support (CDS) model is proposed. The model is widely applicable and scales to realistic, complex decision problems based on adaptive learning. The utility of a decision is discussed and four types of decisions associated with CDS model are identified. The CDS model is designed to learn decision utilities. Data enrichment is introduced to promote the effectiveness of learning. Grouping is introduced for large-scale decision learning. Introspection and adjustment are presented for adaptive learning. Triage recommendation is incorporated to indicate the trustworthiness of suggested decisions.
The CDS model and methodologies are integrated into an architecture using concepts from cognitive computing. The proposed architecture is implemented with an example use case to inventory management.
Reinforcement learning (RL) is discussed as an alternative, generalized adaptive learning engine for the CDS system to handle the complexity of many problems with unknown environments. An adaptive state dimension with context that can increase with newly available information is discussed. Several enhanced components for RL which are critical for complex use cases are integrated. Deep Q networks are embedded with the adaptive learning methodologies and applied to an example supply chain management problem on capacity planning.
A new approach using Ito stochastic processes is proposed as a more generalized method to generate non-stationary demands in various patterns that can be used in decision problems. The proposed method generates demands with varying non-stationary patterns, including trend, cyclical, seasonal, and irregular patterns. Conventional approaches are identified as special cases of the proposed method. Demands are illustrated in realistic settings for various decision models. Various statistical criteria are applied to filter the generated demands. The method is applied to a real-world example.
The CDS model and methodologies are integrated into an architecture using concepts from cognitive computing. The proposed architecture is implemented with an example use case to inventory management.
Reinforcement learning (RL) is discussed as an alternative, generalized adaptive learning engine for the CDS system to handle the complexity of many problems with unknown environments. An adaptive state dimension with context that can increase with newly available information is discussed. Several enhanced components for RL which are critical for complex use cases are integrated. Deep Q networks are embedded with the adaptive learning methodologies and applied to an example supply chain management problem on capacity planning.
A new approach using Ito stochastic processes is proposed as a more generalized method to generate non-stationary demands in various patterns that can be used in decision problems. The proposed method generates demands with varying non-stationary patterns, including trend, cyclical, seasonal, and irregular patterns. Conventional approaches are identified as special cases of the proposed method. Demands are illustrated in realistic settings for various decision models. Various statistical criteria are applied to filter the generated demands. The method is applied to a real-world example.
ContributorsKee, Seho (Author) / Runger, George C. (Thesis advisor) / Escobedo, Adolfo (Committee member) / Gel, Esma (Committee member) / Janakiram, Mani (Committee member) / Rogers, Dale (Committee member) / Arizona State University (Publisher)
Created2020
Description
In today’s rapidly changing world and competitive business environment, firms are challenged to build their production and distribution systems to provide the desired customer service at the lowest possible cost. Designing an optimal supply chain by optimizing supply chain operations and decisions is key to achieving these goals.
In this research, a capacity planning and production scheduling mathematical model for a multi-facility and multiple product supply chain network with significant capital and labor costs is first proposed. This model considers the key levers of capacity configuration at production plants namely, shifts, run rate, down periods, finished goods inventory management and overtime. It suggests a minimum cost plan for meeting medium range demand forecasts that indicates production and inventory levels at plants by time period, the associated manpower plan and outbound shipments over the planning horizon. This dissertation then investigates two model extensions: production flexibility and pricing. In the first extension, the cost and benefits of investing in production flexibility is studied. In the second extension, product pricing decisions are added to the model for demand shaping taking into account price elasticity of demand.
The research develops methodologies to optimize supply chain operations by determining the optimal capacity plan and optimal flows of products among facilities based on a nonlinear mixed integer programming formulation. For large size real life cases the problem is intractable. An alternate formulation and an iterative heuristic algorithm are proposed and tested. The performance and bounds for the heuristic are evaluated. A real life case study in the automotive industry is considered for the implementation of the proposed models. The implementation results illustrate that the proposed method provides valuable insights for assisting the decision making process in the supply chain and provides significant improvement over current practice.
In this research, a capacity planning and production scheduling mathematical model for a multi-facility and multiple product supply chain network with significant capital and labor costs is first proposed. This model considers the key levers of capacity configuration at production plants namely, shifts, run rate, down periods, finished goods inventory management and overtime. It suggests a minimum cost plan for meeting medium range demand forecasts that indicates production and inventory levels at plants by time period, the associated manpower plan and outbound shipments over the planning horizon. This dissertation then investigates two model extensions: production flexibility and pricing. In the first extension, the cost and benefits of investing in production flexibility is studied. In the second extension, product pricing decisions are added to the model for demand shaping taking into account price elasticity of demand.
The research develops methodologies to optimize supply chain operations by determining the optimal capacity plan and optimal flows of products among facilities based on a nonlinear mixed integer programming formulation. For large size real life cases the problem is intractable. An alternate formulation and an iterative heuristic algorithm are proposed and tested. The performance and bounds for the heuristic are evaluated. A real life case study in the automotive industry is considered for the implementation of the proposed models. The implementation results illustrate that the proposed method provides valuable insights for assisting the decision making process in the supply chain and provides significant improvement over current practice.
ContributorsAlmatooq, Nourah (Author) / Askin, Ronald (Thesis advisor) / Sefair, Jorge (Thesis advisor) / Gel, Esma (Committee member) / Pan, Rong (Committee member) / Arizona State University (Publisher)
Created2020
Description
Breeding seeds to include desirable traits (increased yield, drought/temperature resistance, etc.) is a growing and important method of establishing food security. However, besides breeder intuition, few decision-making tools exist that can provide the breeders with credible evidence to make decisions on which seeds to progress to further stages of development. This thesis attempts to create a chance-constrained knapsack optimization model, which the breeder can use to make better decisions about seed progression and help reduce the levels of risk in their selections. The model’s objective is to select seed varieties out of a larger pool of varieties and maximize the average yield of the “knapsack” based on meeting some risk criteria. Two models are created for different cases. First is the risk reduction model which seeks to reduce the risk of getting a bad yield but still maximize the total yield. The second model considers the possibility of adverse environmental effects and seeks to mitigate the negative effects it could have on the total yield. In practice, breeders can use these models to better quantify uncertainty in selecting seed varieties
ContributorsOzcan, Ozkan Meric (Author) / Armbruster, Dieter (Thesis advisor) / Gel, Esma (Thesis advisor) / Sefair, Jorge (Committee member) / Arizona State University (Publisher)
Created2019
Description
Student retention is a critical metric for many universities whose intention is to support student success. The goal of this thesis is to create retention models utilizing machine learning (ML) techniques. The factors explored in this research include only those known during the admissions process. These models have two goals: first, to correctly predict as many non-returning students as possible, while minimizing the number of students who are falsely predicted as non-returning. Next, to identify important features in student retention and provide a practical explanation for a student's decision to no longer persist. The models are then used to provide outreach to students that need more support. The findings of this research indicate that the current top performing model is Adaboost which is able to successfully predict non-returning students with an accuracy of 54 percent.
ContributorsWade, Alexis N (Author) / Gel, Esma (Thesis advisor) / Yan, Hao (Thesis advisor) / Pavlic, Theodore (Committee member) / Arizona State University (Publisher)
Created2021
Description
Complex systems appear when interaction among system components creates emergent behavior that is difficult to be predicted from component properties. The growth of Internet of Things (IoT) and embedded technology has increased complexity across several sectors (e.g., automotive, aerospace, agriculture, city infrastructures, home technologies, healthcare) where the paradigm of cyber-physical systems (CPSs) has become a standard. While CPS enables unprecedented capabilities, it raises new challenges in system design, certification, control, and verification. When optimizing system performance computationally expensive simulation tools are often required, and search algorithms that sequentially interrogate a simulator to learn promising solutions are in great demand. This class of algorithms are black-box optimization techniques. However, the generality that makes black-box optimization desirable also causes computational efficiency difficulties when applied real problems. This thesis focuses on Bayesian optimization, a prominent black-box optimization family, and proposes new principles, translated in implementable algorithms, to scale Bayesian optimization to highly expensive, large scale problems. Four problem contexts are studied and approaches are proposed for practically applying Bayesian optimization concepts, namely: (1) increasing sample efficiency of a highly expensive simulator in the presence of other sources of information, where multi-fidelity optimization is used to leverage complementary information sources; (2) accelerating global optimization in the presence of local searches by avoiding over-exploitation with adaptive restart behavior; (3) scaling optimization to high dimensional input spaces by integrating Game theoretic mechanisms with traditional techniques; (4) accelerating optimization by embedding function structure when the reward function is a minimum of several functions. In the first context this thesis produces two multi-fidelity algorithms, a sample driven and model driven approach, and is implemented to optimize a serial production line; in the second context the Stochastic Optimization with Adaptive Restart (SOAR) framework is produced and analyzed with multiple applications to CPS falsification problems; in the third context the Bayesian optimization with sample fictitious play (BOFiP) algorithm is developed with an implementation in high-dimensional neural network training; in the last problem context the minimum surrogate optimization (MSO) framework is produced and combined with both Bayesian optimization and the SOAR framework with applications in simultaneous falsification of multiple CPS requirements.
ContributorsMathesen, Logan (Author) / Pedrielli, Giulia (Thesis advisor) / Candan, Kasim (Committee member) / Fainekos, Georgios (Committee member) / Gel, Esma (Committee member) / Montgomery, Douglas (Committee member) / Zabinsky, Zelda (Committee member) / Arizona State University (Publisher)
Created2021
Description
The findings of this project show that through the use of principal component analysis and K-Means clustering, NBA players can be algorithmically classified in distinct clusters, representing a player archetype. Individual player data for the 2018-2019 regular season was collected for 150 players, and this included regular per game statistics, such as rebounds, assists, field goals, etc., and advanced statistics, such as usage percentage, win shares, and value over replacement players. The analysis was achieved using the statistical programming language R on the integrated development environment RStudio. The principal component analysis was computed first in order to produce a set of five principal components, which explain roughly 82.20% of the total variance within the player data. These five principal components were then used as the parameters the players were clustered against in the K-Means clustering algorithm implemented in R. It was determined that eight clusters would best represent the groupings of the players, and eight clusters were created with a unique set of players belonging to each one. Each cluster was analyzed based on the players making up the cluster and a player archetype was established to define each of the clusters. The reasoning behind the player archetypes given to each cluster was explained, providing details as to why the players were clustered together and the main data features that influenced the clustering results. Besides two of the clusters, the archetypes were proven to be independent of the player's position. The clustering results can be expanded on in the future to include a larger sample size of players, and it can be used to make inferences regarding NBA roster construction. The clustering can highlight key weaknesses in rosters and show which combinations of player archetypes lead to team success.
ContributorsElam, Mason Matthew (Author) / Armbruster, Dieter (Thesis director) / Gel, Esma (Committee member) / Computer Science and Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2019-05
Description
In the past, Industrial Engineering/Engineering Management Capstone groups have not provided adequate documentation of their project data, results, and conclusions to both the course instructor and their project sponsors. The goal of this project is to mitigate these issues by instituting a knowledge management system with one of ASU’s cloud storage tools, OSF, and by updating course rubrics to reflect knowledge sharing best practices. This project used existing research to employ tactics that promote the long-term use of this system. In addition, data specialists from ASU Library’s Research and Data Management department were involved.
ContributorsWade, Alexis Nicole (Author) / Juarez, Joseph (Thesis director) / Gel, Esma (Committee member) / Industrial, Systems & Operations Engineering Prgm (Contributor) / Barrett, The Honors College (Contributor)
Created2019-12
Description
The premise of this thesis developed from my personal interests and undergraduate educational experiences in both industrial engineering and design studies, particularly those related to product design. My education has stressed the differences in the ways that engineers and designers approach problem solving and creating solutions, but I am most interested in marrying the two mindsets of designers and engineers to better solve problems creatively and efficiently.
This thesis focuses on the recent appearance of generative design technology into the world of industrial design and engineering as it relates to product development. An introduction to generative design discusses the uses and benefits of this tool for both designers and engineers and also addresses the challenges of this technology. The relevance of generative design to the world of product development is discussed as well as the implications of how this technology will change the roles of designers and engineers, and especially their traditional design processes. The remainder of this paper is divided into two elements. The first serves as documentation of my own exploration of using generative design software to solve a product design challenge and my reflections on the benefits and challenges of using this tool. The second element addresses the need for employing quantitiative methodologies within the generative design process to aid designers in selecting the most advantageous design option when presented with generative outcomes. Both sections aim to provide more context to this new design process and seek to answer questions about some of the ambiguous processes of generative design.
This thesis focuses on the recent appearance of generative design technology into the world of industrial design and engineering as it relates to product development. An introduction to generative design discusses the uses and benefits of this tool for both designers and engineers and also addresses the challenges of this technology. The relevance of generative design to the world of product development is discussed as well as the implications of how this technology will change the roles of designers and engineers, and especially their traditional design processes. The remainder of this paper is divided into two elements. The first serves as documentation of my own exploration of using generative design software to solve a product design challenge and my reflections on the benefits and challenges of using this tool. The second element addresses the need for employing quantitiative methodologies within the generative design process to aid designers in selecting the most advantageous design option when presented with generative outcomes. Both sections aim to provide more context to this new design process and seek to answer questions about some of the ambiguous processes of generative design.
ContributorsElgin, Mariah Crystal (Author) / Bacalzo, Dean (Thesis director) / Gel, Esma (Committee member) / Industrial, Systems & Operations Engineering Prgm (Contributor) / Dean, Herberger Institute for Design and the Arts (Contributor) / Barrett, The Honors College (Contributor)
Created2019-05
Description
The Difference Engine at Arizona State University developed the Women’s Power and Influence Index (WPI) in order to combat the systemic inequality faced by women in the workplace. It aims to analyze data, such as Equal Employment Opportunity data, from various Fortune 500 companies to provide a measure of workplace inequality as well as encourage these institutions to adopt more equitable policies. By rating companies based on what truly matters to women, ASU’s Difference Engine hopes to help both women in existing career paths as well as women seeking a new career or position in companies.
However, in order for the WPI to become a relevant scoring metric of gender equality within the workplace, we must raise awareness about the issue of gender equality and of the index itself. By raising awareness about gender inequality as well as inspiring companies to further equality within their workplaces, the WPI will serve to have an integral role in increasing gender equality in the workplace. Our approach for raising awareness utilizes two different strategies: (1) establishing a new version of the WPI website that is both informative and aesthetically pleasing and (2) generating social media content on TikTok that appeal to a variety of audiences and introduce them to the WPI and our mission.
ContributorsThomas, Elisa (Author) / Tieu, Lienna (Co-author) / Howard, Brooke (Co-author) / Zaffar, Ehsan (Thesis director) / Gel, Esma (Committee member) / Barrett, The Honors College (Contributor) / School of Social Transformation (Contributor) / School of Politics and Global Studies (Contributor) / Division of Teacher Preparation (Contributor) / Sanford School of Social and Family Dynamics (Contributor)
Created2022-05
Description
The Difference Engine at Arizona State University developed the Women’s Power and Influence Index (WPI) in order to combat the systemic inequality faced by women in the workplace. It aims to analyze data, such as Equal Employment Opportunity data, from various Fortune 500 companies to provide a measure of workplace inequality as well as encourage these institutions to adopt more equitable policies. By rating companies based on what truly matters to women, ASU’s Difference Engine hopes to help both women in existing career paths as well as women seeking a new career or position in companies. However, in order for the WPI to become a relevant scoring metric of gender equality within the workplace, we must raise awareness about the issue of gender equality and of the index itself. By raising awareness about gender inequality as well as inspiring companies to further equality within their workplaces, the WPI will serve to have an integral role in increasing gender equality in the workplace. Our approach for raising awareness utilizes two different strategies: (1) establishing a new version of the WPI website that is both informative and aesthetically pleasing and (2) generating social media content on TikTok that appeal to a variety of audiences and introduce them to the WPI and our mission.
ContributorsThomas, Elisa (Author) / Tieu, Lienna (Co-author) / Howard, Brooke (Co-author) / Zaffar, Ehsan (Thesis director) / Gel, Esma (Committee member) / Barrett, The Honors College (Contributor) / School of Social Transformation (Contributor)
Created2022-05