Browse Prior Art Database

Dynamic clustering of service technicians and service requests in workforce scheduling

IP.com Disclosure Number: IPCOM000237139D
Publication Date: 2014-Jun-05
Document File: 4 page(s) / 239K

Publishing Venue

The IP.com Prior Art Database

Related People

Dr. Sleman Saliba: INVENTOR [+3]

Abstract

A common practical problem in the workforce scheduling industry is to divide the customer base intelligently into smaller areas and to assign the service technicians into these areas. The known static geographic clustering approach is extended by two innovations. Firstly, the clusters or dispatch areas are not static but are dynamically adapted. This adaption can be done automatically on a regular basis. In this way, fast reaction to changing environments becomes possible. Secondly, more factors and criteria for the clustering are included, like skills of service technicians, amount of service requests and service technicians as well as urgency level of the affected requests

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 29% of the total text.

Dynamic clustering of service technicians and service requests in workforce scheduling

Background

The task of workforce optimization in general solves the problem of assigning a set of customer orders to a set of technicians. This assignment consists of many constraints such as customer order time windows, limited working hours for technicians, required break times and break durations for technicians, technical skills of the technicians, possible pre-assignment of technicians into shifts, and travel time constraints between any subsequent customer orders.

Service orga­nizations in the utility and commu­nications industries operate in an ever-changing environment of rising costs, complex regulations, merg­ers and acquisitions, and custom­ers’ high expectations of reliability, responsiveness and service quality. To meet these challenges, it is critical that utilities optimize their service delivery by making efficient use of field technicians and equipment, and improve the reli­ability of critical assets.

Beside the general problem and the constraints at hand, different optimization objectives can be used for workforce optimization. Typical objectives are the maximization of overall executed orders per time interval, the minimization of violated customer time windows, and the minimization of travel time.

The workforce optimization problem needs to be solved at different times of the day. Usually, a more time consuming and rather complex optimization is performed during the night hours such that a good solution for the next day is generated to be used as a baseline for the guidance of the technicians. This solution is often called day-ahead or offline solution.

Based on this day-ahead solution, the technicians receive their roster in the morning when they call in for duty. The duty-roster or service schedule can be distributed in paper form or via mobile electronic devices, such as PDAs. The human operator supervising the system uses the pre-calculated plan for guidance.

For the offline solution of the workforce scheduling problem, the column generation algorithm with labeling approach for solving instances of about 100 shifts and 500 orders has been developed. This is the desired size of a dispatcher working place. However, customers usually deal with a much larger customer base. In the biggest installations, the up to 10,000 shifts and 100,000 orders need to be handled.

A common problem is to divide the customer base intelligently into smaller areas that are manageable by a human dispatcher. Currently this division is done once and statically. This means that at each new installation the customer base is divided into geographic areas, which are then clustered into dispatch areas that are supervised by dispatchers. Throughout this description, a cluster is the set of service technicians and service requests that are assigned to a dispatch area. Traditionally, this pre-defined clustering is only changed manually in case of heavy resource...