A-Plan

Hide titles
Aim: 

Assignment of staff to work tasks is a complex problem that involves a large number of factors and requires a lot of expertise. Long term as well as short term requirements need to be met which demands flexible solutions. Software tools can aid planners in reaching optimal dispatching plans but currently available solutions provide only incomplete support. This thesis describes the design, development, and evaluation of a prototype for semi-automated assignment planning called A-Plan. We have carried out this work in the context of a gas device maintenance provider. In A-Plan, assignments of service technicians to customers are displayed visually and can be modified by direct manipulation. Smooth cooperative work is possible and an optimization algorithm has been integrated that facilitates semi-automatic planning. A qualitative evaluation with potential users and IT professionals provided encouraging feedback on the proposed integration of automated methods and interactive visual interfaces.

Team: 
  • Thomas Schneider, Vienna University of Technology, Institute of Software Technology and Interactive System, Vienna, Austria 
  • Wolfgang Aigner, Vienna University of Technology, Institute of Software Technology and Interactive System, Vienna, Austria
  • Silvia Miksch, Vienna University of Technology, Institute of Software Technology and Interactive System, Vienna, Austria 

User and task analysis

To illustrate the problem, all tasks and the currently used software were investigated. We found many usability problems in the existing software system (SAP CRM). The process of an incoming call is illustrated here:

  1. Customers report their requests through the company’s call center. During the customer call, the dispatcher has to arrange a target date with the customer.
  2. During contextual observation we encountered that in many cases the dispatcher does not use the software at all. She has to answer questions and confirm dates of customers on the phone within seconds, and this is not possible with the mentioned software. Therefore, the users of the software came up with a workaround that makes rapid answering on the phone possible. Rather than retrieving information from the electronic system, two paper-based artifacts are used.
  3. In the evening all data is recorded from the paper based artefacts to the current system. Consequence is a high error rate and more effort. 
  4. The data is replicated to the mobile application of the service technicians.
  5. The technician does the repair or maintenance of the device.

The goal for this user & task analysis besides gaining more knowledge about the domain, user requirements, and desires, was the creation of representative user profiles, their goals, and the construction of interaction scenarios based on this user model. Following the user-centered design approach by Cooper [2], this lead to the creation of scenarios and personas that aid design and evaluation.

Design

Two guiding lines of the design of A-Plan were to fulfill user requirements and to avoid reported problems and issues of current work practice. To support and ease the worflow of users, an automated planning function should be integrated into the software. Furthermore, multiple users should be able to work simultaneously with the data while being aware of each others' actions. Following Cooper et al.'s recommendation [3] we eliminated save buttons and avoided OK buttons. Instead, every action should be saved automatically and an undo function should be available for the user to take back unwanted operations. Following that, every transaction should be saved immediately and distributed to the other clients.

The screen is divided into three areas (see image below):

  • The planning area is the place where the user can view, insert, move and delete assignments.
  • Details are shown on the right side of the window. In this location the user can also start actions like searching for a customer and planning of assignments.
  • Collaboration: In the lower area of the window the user is supported in the cooperation with other users.

A-Plan integrates interactive visualizations with automated planning and supports collaborative work. It uses an interactive visualization for presenting planned assignments that is based on timelines. A detail view of assignments includes an interactive map view for localizing customers. Furthermore, an automated planning functionality based on the savings algorithm has been integrated which allows for bulk planning of recurring service contracts and is supported by a heatmap visualization. The algorithm accounts for a complex set of constraints like geographic areas and timings and is suggesting an automatically optimized set of assignments. The suggested plan can be reviewed and altered by the user via the interactive visual interface.

Conclusion

The main contribution of our work is that we have demonstrated the successful application of a Visual Analytics approach in the context of resource scheduling. We have shown an effective integration of automatic methods and interactive visualizations based on a user-centric development approach. The integration of the strength of both the human and the computer enables the creation a powerful environment for a set of non-trivial and complex tasks.

 

Papers

Thomas Schneider, Wolfgang Aigner: A-Plan: Integrating Interactive Visualization With Automated Planning for Cooperative Resource Scheduling. In Proc. to the TAVA Special Track on Theory and Applications of Visual Analytics at the 11th International Conference on Knowledge Management and Knowledge Technologies, Graz, Austria, ACM, New York, September, 2011, http://dx.doi.org/10.1145/2024288.2024341

Thomas Schneider: A Cooperative Resource Scheduling Tool With Automated Planning; Master Thesis. Supervisors: S. Miksch, W. Aigner; Vienna University of Technology,Institute of Software Technology and Interactive Systems, Vienna, Austria, September 2011. 

References

[1] G. Clarke and J. V. Wright. Scheduling of Vehicles from a Central Depot to a Number of Delivery Points. 
     Oper. Res., 12(4):568-581, 1964. 
[2] A. Cooper. The Inmates Are Running The Asylum. Sams Publishing, 2004.
[3] A. Cooper, R. Reimann, and D. Cronin. About Face 3: The Essentials of Interaction Design. Wiley Publishing, 2007.

Contact: 

Thomas Schneider

Status: 
finished
AttachmentSize
APlan.zip3.16 MB