With the abundance of available data, many enterprises seek to implement data-driven prescriptive analytics to help them make informed decisions. These prescriptive policies need to satisfy operational constraints, and proactively eliminate rule conflicts, both of which are ubiquitous in practice. It is also desirable for them to be simple and interpretable, so they can be easily verified and implemented. Existing approaches from the literature center around constructing variants of prescriptive decision trees to generate interpretable policies. However, none of the existing methods is able to handle constraints. In this paper, we propose a scalable method that solves the constrained prescriptive policy generation problem. We introduce a novel path-based mixed-integer program (MIP) formulation which identifies a (near) optimal policy efficiently via column generation. The policy generated can be represented as a multiway-split tree which is more interpretable and informative than binary-split trees due to its shorter rules. We demonstrate the efficacy of our method with extensive computational experiments on both synthetic and real datasets.

This paper presents design rule checking (DRC) methods to address challenges caused by non-Manhattan geometries that are widely present in photonic integrated circuit (PIC) physical designs. Verifying such layouts with traditional DRC tools results in a huge number of false errors that are impossible to debug; and some physical rules are simply uncheckable. We demonstrate the use of an extended DRC technique on photonic designs to tackle these issues. The results show that the technique is well-suited for silicon photonics DRC, enabling false errors to be filtered in a controllable manner, leading to easier debugging with layout annotation; multi-dimensional rule checking to be performed; and mathematical expressions to be used for more accurate descriptions of physical rules.

The classical definition of knowledge management promises to get the right knowledge to the right people at the right time so they can make the best decision [G. Petrash, 1996]. Autonomic systems, on the other hand, are expected to find and apply the right knowledge for self-managing purposes without human intervention. This article discusses the components to be built around a system to enable self-healing and managing capabilities. These are defined and described in this article as self-knowledge, self-monitoring, self-learning, problem detection, diagnosis, and search and solution components. Interaction of these system components to make knowledge available for self-healing purposes is also discussed.

In the e-commerce market, we identified the need for an intelligent system that has the capacity to keep the customers well-informed. One form of information capture and discovery is through the services of event notification equipped with a prediction capability. With such a capability, an e-commerce participant can leverage (near-)real-time prediction regarding almost every area in the e-commerce market, ranging from stock selection to market trends. This paper describes PENS (Predictive Event Notification System). This system is part of an effort to enrich our e-commerce platform. PENS is presented from two different perspectives: the event notification service and the event prediction service. PENS has been applied to the domain of product support services and has generated satisfactory results for customers. A high-level description of PENS' architecture and its implementation is given in this paper.