Siebel Field Service Guide > Setting Up and Using Scheduling >

About Scheduling Heuristic Methods

Both the ABS and the Optimizer use heuristic methods to schedule appointments. These heuristic methods are specified as part of the parameter set configuration—see Creating Parameter Sets for Scheduling for more information.

Heuristics for the Appointment Booking System

There are three possible heuristic methods that the ABS can use to schedule appointments:

• Earliest First. The earliest available time slots are included in the set. The ABS examines each person in the schedule to find the first open slots after the Earliest Start date. This method may provide the fastest possible response for the activity. Companies that are booking activities for the same day prefer this method because it maximizes the amount of work people are performing.
• Resource Loading. Time slots are selected so that the schedules for employees are filled to the extent possible, one employee at a time. If two or more qualified employees are available, the time slots for the employee with the most complete (filled-up) schedule are selected for inclusion in the set. This heuristic method is primarily used for scheduling partners or on-call resources where the company wants to use as few resources as possible to meet the service needs.

  TIP:   The number of time slots returned in response to the request is configurable through the parameter ABS - Default Number of Slots.

• Resource Leveling. Time slots are selected so that the schedules for qualified employees are evenly loaded, to the extent possible. This heuristic method is useful in situations where companies have salaried field service engineers who should be used when on schedule.

  NOTE:  Selecting this heuristic method does not mean that the time slots for all employees in the schedule will be displayed. Multiple slots from the same employee might be shown if that employee is more lightly loaded.

Heuristics for the Optimizer

The Optimizer provides a choice of methods for obtaining solutions for a service region's schedule. Each of these methods involves a different strategy to improve the schedule. All methods use operations to move two or a few appointments and activities in a schedule and then verify improvements in the overall cost of the schedule. For example, the Optimizer may try to swap a pair of activities between two field service engineers. This can change the cost of the schedule by changing the cost of travel, the amount and cost of overtime, and the rate billed by the field service engineer. Any reduction in cost is an improvement in the schedule. For more information, see Defining Cost Functions for the Optimizer.

There are five different classes of operations for moving appointments and activities in a schedule. Each optimization method uses all of these scheduling heuristics.

There are two basic optimization methods, which can be used alone, and three methods that can be used in combination with the basic methods. The following basic methods consistently search for and accept only improvements (lower cost) in the schedule:

• Greedy search. This method starts with an existing schedule and finds the first move that improves the schedule, accepts the move, and then uses this solution to find the next improvement. A Greedy search repeats this process until there is no more improvement or it reaches a time limit. This method is relatively fast, but the result solution is not as good as other methods.
• Steepest search. This method starts with an existing schedule, tries all moves, accepts the move that provides the greatest improvement in the schedule, and then uses this solution to find the next improvement. A Steepest search repeats this process until there is no more improvement or it reaches a time limit. This method takes longer, but generally produces lower-cost schedules.

The following methods, combined with either the Greedy search or the Steepest search, allow moves that temporarily increase the cost of a schedule to arrive at significant, overall improvements in the schedule. In all cases, the Greedy search or the Steepest search quickly finds an improved schedule, and then one of the following methods takes over and searches for improvements:

• Tabu search. This method accepts the next best solution even if it is not an improvement over the previous schedule. It keeps a tabu list of finite length that contains the results of previous moves. The Optimizer cannot repeat a move until a move drops off the list.
• Fast Guided Local search. This method adjusts the cost of a solution to reflect the number of times the Optimizer tried a move. This allows the Optimizer to try a wider range of changes.
• Fast Guided Tabu search. This method combines the Tabu search and the Fast Guided Local search. It often finds good solutions faster than either the Tabu or Fast Guided Local searches.