This is a novel graphical interface for event scheduling (examination timetabling).  The evaluations of the system showed that novices could quickly learn to use the system and were able to generate solutions of a quality to match commercial strength automated scheduling systems.  The design of the interface reveals the underlying structure of the task domain to users, with the consequence that they can use sophisticated recursive strategies rather than the trial-and-error approaches observed with conventional graphical interfaces.  Yellow exam icons are allocated to blue room/time slots.  All horizontal dimensions are temporal properties (day, time and durations of slots and exams) and all vertical dimensions are spatial properties (rooms, seating capacity, exam size).  Violating exams are shown by connecting lines between the icons.  The red icon reveals all the exams taken by anyone in the selected white exam (centre), which enables the user so easily see where it may be re-allocated to eliminate conflicts without causing new violations. 

Cheng, P. C.-H., Barone, R., Cowling, P. I., & Ahmadi, S. (2002). Opening the information bottleneck in complex scheduling problems with a novel representation: STARK diagrams. In M. Hegarty, B. Meyer & N. H. Narayanan (Eds.), Diagrammatic representations and inference: Second International Conference, Diagrams 2002 (pp. 264-278). Berlin: Springer-Verlag.
Cowling, P. I., Ahmadi, S., Cheng, P. C.-H., & Barone, R. (2002). Combining human and machine intelligence to produce effective examination timetables. In  L.Wang, K. Tan,  C. Furuhashi., J-H  Kim, and X. Yao (Eds.), Proceedings of the 4th Asia-Pacific Conference on Simulated Evolution And Learning (SEAL2002), Singapore; (pp 662-666). (ISBN 981-04-7523-3).

This is a novel graphical interface for personnel scheduling (nurse rostering).  The evaluations of the system showed that novices could quickly learn to use the system and were able to generate solutions of a quality to match an expert scheduler.  The design of the interface reveals the underlying structure of the problem constraints to users.  White icons show constraints that are more than satisfied, grey icons show constraints that are just satisfied and black icons are unsatisfied constraints.  Hence, the solution can be improved by finding areas that are black (e.g., centre 1/3 from the bottom) and moving icons to areas that are white (e.g., the column 1/3 from the left). 

Barone, R., & Cheng, P. C.-H. (2004). Representations for problem solving: on the benefits of integrated structure. In E. Banissi, K. Börner, C. Chen, et al. (Eds.), Proceedings of the 8th International Conference on Information Visualisation (pp. 575-580). Los Alamitos, CA: IEEE.

ROLLOUT was designed as a system for diagrammatic production planning and scheduling in factory-scale and in-store bakeries.  The new representations coherently encode all the information needed for the planning of orders and the scheduling of production.  ROLLOUT diagrams reveal critical relations as unique easily recognizable patterns.  Experienced bakery managers are able to see and exploit these high-order regularities in their problem solving, which were previously hidden.  Novices are able to create good schedules after a minimal amount of training.  The top half the diagram supports the decomposition of orders (grey stacks) into to coordinated batches of production of like products.  The bottom supports the allocation of production runs (staircases) for batches of particular product types.  Violations of the plan, schedule and the physical plant constraints are highlighted in red and can be simply resolved by juggling the icons. 

Cheng, P. C.-H., & Barone, R. (2007). Representing complex problems: A representational epistemic approach. In D. H. Jonassen (Ed.), Learning to solve complex scientific problems. (pp. 97-130). Mahmah, N.J.: Lawrence Erlbaum Associates.

These systems were created in order to explore methods by which human knowledge and flexibility could be combined with the computational power and information storage capabilities of computers, so that each could compensate for the limitations of the other.  Both address the problem of examination timetabling. 
    The VAST system (top) successfully used automatic pre-clustering of the raw data followed by human guided allocation and refined processes. By pre-clustering events, large numbers of constraints are automatically satisfied, which means that problem solvers can focus their attention on tricky constraints that would not normally be handled by standard evaluation functions and hence not addressed by an automated algorithm.
    The KNIGHT system (bottom) demonstrated the feasibility of allowing humans to guide TABU optimization. To influence the potential of automated reallocation of events users could steer the algorithm in directions they considered to be potentially fruitful. They did this by inspecting visualizations of the solution process, using their own particular knowledge of the problem, and then varying the mobility of different classes of events so that some would be treated preferentially.

Ranson, D., & Cheng, P. C.-H. (2008). VAST Improvements to Diagrammatic Scheduling using Representational Epistemic Interface Design. In G. Stapleton, J. Howse & J. Lee (Eds.), Diagrammatic Representation and Inference, Lecture Notes in Aritificial Intelligence 5223 (pp. 141-155). Berlin: Springer
Ranson, D. (2008).  Interactive Visualisations to Improve Automated Scheduling Systems.  PhD Thesis.  (Contact Peter Cheng for a copy.)