Practical Implications of Preference-Based Conference Scheduling Conference scheduling involves organizing presentations into sessions which are assigned to time periods and rooms. This includes assuring a feasible utilization of time periods and rooms, and avoiding individual schedule conflicts. The problem increases in complexity by considering the preferences of presenters: for time periods, for rooms, etc. A greater level of complexity comes from considering the preferences of conference attendees, which we define as preference-based conference scheduling (PBCS). This article provides a structure on which PBCS problems are founded, including empirical demonstration of solution techniques. In addition, real-world strategic planning issues of flexibility and preference detail are explored. Key words: scheduling; optimization; service operations; services Submissions and Acceptance: Received June 2003; revision received October 2003; accepted December 2003. 1. Introduction Computer technology has impacted the delivery of academic conferences in only minor ways. Presenters use computer projectors to display presentation slides, yet they are pretty much the same slides that would otherwise have been printed on transparency film and shown on an overhead projector. The meetings are still physical, despite some potential for electronic, virtual meetings. That physical presence seems to be a major advantage of conferences over electronic research interactions. Even though the delivery of conferences is littleimpacted by computer technology, the planning of conferences is being revolutionized. Conference planning has traditionally been so arduous of a task that planners were happy with any feasible conference schedule that could be derived. Now, Internet and computer technology hold the potential to allow vastly more complex conference planning processes. Hundreds, even thousands, of conference participants can participate in the planning process by helping specify a conference schedule that is not only feasible, but is optimized (as defined by a chosen objective function).
This article lays out theory behind preference-based conference scheduling (PBCS), which is scheduling to maximize the satisfaction of various preferences (as well as meet various feasibility constraints). Conference scheduling has not received a great deal of attention in prior research literature, although we are able to draw from research in related areas. There are issues of conference scheduling that are uniquely interesting, which will be discussed. At a basic level, the conference scheduling problem involves assigning sessions to time periods and rooms. The overriding requirements are that (1) no more than one session is in a given room during a given time period, and (2) no presenter can be required to be in more than one place at a given time. These requirements simply define solution feasibility. Under normal circumstances, a feasible solution that meets both requirements would exist. Such a problem has no objective function, thus nothing to optimize. As a result, conference scheduling becomes a relatively uninteresting problem of identifying a feasible solution. If the conference scheduling problem were fitted with an objective function, what would the objective function look like? This article considers the possibility of a customer-satisfaction objective function. By defining the "customers" as those attending the conference, the objective might be to provide the best possible conference schedule for everyone attending the conference. As will be expounded later, this implies that people can attend the sessions they want; 1.e., the sessions are held at times that are agreeable and there is sufficient seating capacity. Satisfying participant preferences is the focus of "preference-based" conference scheduling (PBCS). This article outlines the theoretical and practical foundation of the PBCS problem. The context of the problem is described, and it is compared to related problems covered in research literature. The general specification of the problem is outlined, and strategic decisions are discussed. A mathematical formulation is given and the PBCS approach is tested empirically. A final section summarizes and points to future research opportunities. 2. The Conference Scheduling Problem A conference is "a meeting of two or more persons for discussing matters of common concern" (Merriam-Webster 1998). Academic conferences typically involve casual networking, hallway discussions, and other interpersonal dialog. Though valuable, organizing these types of informal meetings is not the primary burden of the conference planner. Conferences are also largely composed of presentations in which presenters describe some paper or topic to an
audience of conference participants. The complex and arduous process of conference scheduling largely involves organizing these presentations into sessions at specific times and in specific rooms. Why is conference scheduling such a complex and arduous process? One reason is that presenters often give multiple presentations during the course of a conference. It is essential that no presenter be required to attend more than one session at a given time. The problem is more complicated if both major and minor "presenters" are considered, a "major presenter" being the person actually speaking to an audience, and a "minor presenter" being a co-author or someone else who needs to be present in the audience in case questions are asked of them. The problem becomes more challenging with the introduction of other "quasi-presenters" such as track chairs and discussants.