An Interactive System for Hiring and Managing Graduate

An Interactive System for Hiring and Managing Graduate Teaching Assistants at the Department of Computer Science and Engineering http: //csce. unl. edu/~gta/ Ryan Lim and Berthe Y. Choueiry Constraint Systems Laboratory • Department of Computer Science and Engineering, University of Nebraska-Lincoln • (rlim|choueiry)@cse. unl. edu Motivation and Goal of the GTA Project At the beginning of each semester, CSE hires a number of Graduate Teaching Assistants (GTAs) as graders, lab supervisors, and instructors. The department needs to assign the GTAs according to their abilities and availability to execute these academic tasks. We believe that a motivated GTA serves the class better, which is beneficial to the students themselves and the faculty teaching the course. Typically, there are 70 different academic tasks each semester and less than 35 GTAs are hired. The problem is often very tight, and the constraints are too numerous for a human to keep track of. In an effort to improve the quality of our program and streamline the managerial operations of our department, we are building a system that enables students to input their academic record and their teaching preferences while providing a human manager with both interactive and automated problem solving capabilities. This poster describes the graphical user interfaces. The algorithmic aspects are described elsewhere. Manager Web front-end The manager login’s is secure. He/She is offered the following options: • Access, edit, and print the data of a particular student applicant (Figure 6) • Access, edit, and print the data of the courses offered (Figure 8) • Hire or remove GTAs and specify their working load (Figure 7) • Assign interactively GTAs to courses • Launch and monitor problem solving by search An interactive system The front-end of the system consists of two web interfaces for data input: one for the students and the other for the manager. The back-end consists of the search algorithms, which are controllable by the manager’s front-end. A My. SQL databased links these three components. GTA web front-end Figure 6 Figure 7 Manager web front-end My. SQL database Search (Lisp) back-end Figure 1 The front-end consists of a web based interface where the GTA and department could logon to change and view their settings. Web front-end Figure 8 Search (Lisp) back-end GTA Web front-end The system links the manager’s interface to a portfolio of search engines being developed by the Constraint System Laboratory. These encompass systematic backtrack heuristic search (implemented), random systematic search (on-going), heurisctic local search (implemented), multi-agent based search (implemented), market-based search (on-going). A candidate TA can securely register on the system anywhere from the Internet. Once logged in, the TA can enter his/her academic record (Figure 4) and edit his/her course selection and teaching preference (Figure 5): The search is performed on the data set obtained from the GTA in the database and is fed to the search engines written in Lisp. Has two components: the GTA (Figures. 2, 3, 4, 5) front-end and the manager front-ends to register and login (Figures. 6, 7, 8). Integration of the back-end and front-end This system has 3 main components. Currently, GTAs are able to register and set their preferences. The manager’s access is partially complete. The search components have not yet been integrated but are operating as a stand-alone module. Future work Figure 2 Figure 3 • Complete the integration of the 3 components. • Enrich the manager’s interface with a capability to run various search engines in parallel and allow them to communicate and cooperate. References 1. R. Glaubius and B. Y. Choueiry, Constraint Modeling and Reformulation in the Context of Academic Task Assignment. In Working Notes of the Workshop Modeling and Solving Problems with Constraints, ECAI 2002. 2. H. Zou. MS Thesis. Con. Syst. Lab, CSE-UNL. In preparation. 3. H. Zou and B. Y. Choueiry, Multi-agent Based Search versus Local Search and Backtrack Search for Solving Tight CSPs: A Practical Case Study. Submitted. 4. V. P. Gudetti. MS Thesis. Con. Syst. Lab, CSE-UNL. In preparation. Figure 4 Figure 5 At any point, the student can examine course descriptions and fill out a survey on the flexibility and ease of the interface. This work is supported by NSF grant #EPS-0091900, the Department of Computer Science and Engineering, and the Constraint Systems Laboratory.