|
9 |
|
|
Description Traditional approaches to contact center scheduling have utilized Erlang and simple queuing models to approximate the real contact center system. Unsatisfactory service levels and high costs often result from the use of these traditional techniques. The reason that these approaches fail to produce efficient solutions is that the underlying models are poor representations of the true system. Discrete event simulation is a powerful modeling tool that allows one to model contact center operations in an extremely realistic fashion. The model can be created to actually mirror the true system. Given a contact center simulation model, a manager can try various schedules and accurately evaluate the resulting contact center performance. In theory, the manager could continue to try schedules until an optimal solution is found. The difficulty with this approach is that the scope of the problem makes human solution techniques ineffective. Consider a contact center that schedules five skill levels every thirty minutes throughout the day. The manager is attempting to determine 5*24*2 = 240 quantities. People can often mentally grasp optimization problems that contain a few variables but a problem with 240 variables is completely intractable for a human. The answer to this dilemma is to use an automated optimization system that can effectively address problems of this size. We propose that a simulation optimization system be developed. Algorithmic development and scenario testing will be involved. Important features will be to integrate optimization with simulation in a manner that:
(1) adequately deals with the inherent uncertainties and routing complexities
reflected in contact center scheduling |
|
|
10 |
|
|
Description Digital optical systems such as optical communication networks and optical computers process digital information, which is optically encoded. These systems have the inherent advantages of speed, parallelism, and immunity to electromagnetic interference (EMI). These systems are characterized by device technology, architectural structure, and digital signal encoding method. Device technology can be optical or a hybrid-optical technology such as optoelectronics. The system architecture can be guided-wave (e.g. fiber optic) or free space. Digital signals can be optically encoded by intensity, polarization, and/or frequency. With such a variety of system design options, coupled with high data transfer rates (speed) and a high degree of parallelism, it is difficult to analyze, quantify, and predict the effects a change in technology, architecture, or methodology will have on the function, the behavior, and the performance of a system. Digital optical systems have evolved in size and in complexity to the extent where current design tools lack the ability to effectively model, simulate, and analyze their behavior. The major functional issues in digital optical systems are optical alignment, power budgeting, timing, synchronization, and control. Timing behavior is defined as the effect the location of a particular signal has on the functioning of a system. Optical signals propage linearly at a constant speed. Signal degradation, the propagation path length, delays along the propagation path, and the duration of the asserted value can cause variations in the behavior of the system. Synchronization is defined as the precise timing relationship between signals. The control behavior of a system is the order of occurrence of synchronized groups of signals. The research in progress models digital optical systems as a timed-colored Petri net (TCPN). A Petri net is a graphical and mathematical modeling method for describing and studying information systems characterized as being concurrent, asynchronous, distributed, parallel, nondeterministic, and/or stochastic. Petri nets are discrete event simulators. For research purposes, the generic Petri net is adapted to modeling digital optical devices and systems by adding timing capabilities and "color" constructs for modeling high-level system properties such as optical signals. The TCPN expresses the structural, functional, and behavioral features of a digital optical system. The research goal is to develop system
simulation and system analysis tools that will: |
|
|
11 |
|
|
Description The Unified Modeling Language (UML) is a language for specifying, visualizing, constructing, and documenting the artifacts of software systems. The UML uses several views and diagrams, such as design class diagrams and collaboration diagrams. We are interested in defining and assessing test adequacy criteria that are based on elements of these diagrams. The criteria will be used to evaluate the adequacy of quality assessment during both design and implementation phases. The proposed work will also develop a method for deriving test suites that satisfy the criteria. |
|
|
12 |
|
|
Description Organizations are struggling to understand ways to effectively and efficiently train and educate employees. Employers want employees with both technical skills and skills that are "socially driven," such as project management, leadership, team participation, etc. As one thinks about this topic in a corporate context, the issues in corporate education and training parallel those found in distance education. What is the best way for technology to support learning and training? The purpose of this research study is to identify key components for organizations to use with their distance-based training and its derivatives (i.e. CBT) to their best advantage. For the purpose of this discussion, distance education is defined as creating a learning environment that facilitates structured learning without the traditional practice of face-to-face interaction in an on-campus environment. This means that computer-based training CDs, video supported training, as well as, Internet-based distance classes meet the spirit of the definition. However, in today's world, distance education usually implies some sort of technological support through the Internet, email or videoconferencing. Ultimately, this leads one to consider how technology and people interact. Overall, research is mixed on the effectiveness of distance learning. Some findings show no difference between groups in distance learning environments while other research show some dysfunctions. However, learning seems to have a significant social component and this aspect of learning has not been incorporated into much of the distance learning models. This project will look at the impact of adding more interaction to the distance environment. For example: Do structured debates on a listserve provide better learning and retention of topics? Do voice-supported materials help students learn? What computer interaction techniques seem to work best to support training and learning? |
|