The number of service visits of Alzheimer’s disease (AD) patients is different from each other and their visit time intervals are non-uniform. Although the literature has revealed many approaches in disease progression modeling, they fail to leverage these time-relevant part of patients’ medical records in predicting disease’s future status. This paper investigates how to predict the AD progression for a patient’s next medical visit through leveraging heterogeneous medical data. Data provided by the National Alzheimer’s Coordinating Center includes 5432 patients with probable AD from August 31, 2005 to May 25, 2017. Long short-term memory recurrent neural networks (RNN) are adopted. The approach relies on an enhanced “many-to-one” RNN architecture to support the shift of time steps. Hence, the approach can deal with patients’ various numbers of visits and uneven time intervals. The results show that the proposed approach can be utilized to predict patients’ AD progressions on their next visits with over 99% accuracy, significantly outperforming classic baseline methods. This study confirms that RNN can effectively solve the AD progression prediction problem by fully leveraging the inherent temporal and medical patterns derived from patients’ historical visits. More promisingly, the approach can be customarily applied to other chronic disease progression problems.

ervice is broadly considered as an application of specialized knowledge, skill, and experience, performed for co-creation of respective values of both consumer and provider. Services are engineered and delivered through a heterogeneous service system. Compared to physical goods in manufacturing, resources, largely people (end users as the service consumer and employees as the service provider) -the main focus of a service system, cannot be held and are more complex to model and manage as people participating in service production and consumption have physiological and psychological issues, cognitive capability, and sociological constraints, etc. As the world becomes more complex and uncertain socially and economically, this research proposes a computational thinking approach to modeling of the dynamics and adaptiveness of a service system, aimed at fully leveraging today's ubiquitous digitalized information, computing capability and computational power so that the service system can be studied qualitatively and quantitatively. Ultimately, with this foundation we will successively and successfully develop the following mechanisms to implement and enhance service systems:• A mechanism to timely capture end users' requirements, changes, expectation and satisfaction in a variety of technical, social, and cultural aspects; • A mechanism to efficiently and cost-effectively provide employees right means and assistances to engineer services while promptly responding the changes; • A mechanism to allow involved people consciously infuse as much intelligence as possible into all levels and aspects of decision-making to assure necessary system adaptiveness for smarter operations.

No abstract

No abstract

No abstract

No abstract

No abstract

Two basic measures, model complexity and model construction efficiency, are usually used to evaluate the implementability (or ease of use in practice) of a methodology for modeling the control of a discrete event manufacturing system (DEMS) on the shop floor. Many well-recognized methods are used to represent and analyze the dynamics of DEMS's, but not many relevant applications have been found in developing control software for the shop floor due to their shortcomings in satisfying these two measures. This paper explores a methodology for modeling the control of a DEMS, which leads to ease of control software development, rather than a new representational/analytical tool, by significantly reducing the model complexity (in terms of the number of required control states) and improving the model construction efficiency. First, an extended finite machine, called a deterministic finite capacity machine (DFCM) with parallel computing capability is developed. Based on DFCM's, the complexity growth function of a DEMS control model is linear in the number of synthesized control components. Then, an automaton structure of a DFCM control model, called structured adaptive supervisory control (SASC), is developed. By referring to supervisory control theory, an SASC model is created with three function layers: acceptance, adaptive supervision, and execution. The well-defined structure ensures that the control model can be constructed systematically.