Title:
Quantum Computation Inspired Constraint Handling and Intelligent Learning for Data-Driven Combinatorial Optimal Decision-Making in Operations Engineering
Quantum Computation Inspired Constraint Handling and Intelligent Learning for Data-Driven Combinatorial Optimal Decision-Making in Operations Engineering
Author(s)
Zou, Pan
Advisor(s)
Jiao, Jianxin (Roger)
Editor(s)
Collections
Supplementary to
Permanent Link
Abstract
Combinatorial optimal decisions are widely observed in operations engineering applications. The emerging field of quantum inspired computation has shown some salient features of quantum information processing that are promised to excel in quantum speedup contributing to solving operational optimization problems of practical relevance more efficiently than conventional paradigms.
This research investigates the potential of quantum inspired optimal decision-making modeling and solutions with the objective of creating mathematical and computational models to advance operations engineering problem solving with respect to three particular technical focuses: (1) Improving computational efficiency through quantum inspired combinatorial optimization; (2) Enhancing domain problem context representation in operations system modeling via quantum entanglement inspired hard constraint handling; and (3) Bolstering data-driven decision making by quantum inspired neural network modeling.
The research proposes and validates a variety of new methods to synthesize a novel quantum inspired decision-making framework, including a twofold update quantum inspired genetic algorithm (TU-QIGA) for computationally efficient combinatorial optimization, a quantum entanglement inspired hard constraint handling approach (QEI-GA) and its variant lightweight quantum inspired genetic algorithm (LQIGA) for optimization with context related hard constraints, and a hybrid quantum inspired semantic neural network (HQISNN) model for learning and supporting data-driven decision making. The accuracy and reliability of the proposed quantum inspired methods and algorithms are assessed based on computational experiments, along with practical applications of industrial case studies.
Sponsor
Date Issued
2022-08-02
Extent
Resource Type
Text
Resource Subtype
Dissertation