Assumptions

We assume the following:

The system operates seven days per week.
There are three shifts per day; day-shift (D), evening-shift (E), and night-shift (N). The week-end days are the last two days.
Each shift is composed of a number of tasks. Each task requires a number of labor with different capabilities.
Each labor will work five shifts per week.

Each labor will have two days off per week, not necessarily to be consecutive.
The system contains different types of labor, classified hierarchically by their capabilities or experiences. A higher qualified labor can substitute a lower qualified one but not vice versa.
The schedule is rotating and is done for one week.
No employee absenteeism.

The proposed knowledge-based system (KBS)

We establish a knowledge-based system to solve the manpower scheduling problem with the same predescribed conditions.
The system is built using Visual Prolog V5.2.
The obtained schedule satisfies all the applicable constraints.

The input data is the tasks requirements from each labor type for each shift.
These requirements are summed to form the shift requirements from each labor type.

The input data of the problem is transformed to facts.
There are other facts that represent the status of the labor for each shift (on- off), the summation of his working shifts, shifts that contain surplus labor… etc.

This step is the proposed algorithm which starts with calculating the workforce size, then assigns on-off shifts and a specific task simultaneously to each labor.
In each assignment iteration, the system searches for a suitable labor which has the minimum workload and assign him a task to do. So, we gurantte that the workload is distributed uniformly among the available labor.
After obtaining the weekly schedule the system distributes the surplus labor on shifts in even manner.

Results are printed in a tabular format showing the required number from each labor type, the on-off shift schedule including the task assignment for each employee.
The table below shows the tour schedule obtained from the KBS

The resulted number of labor from applying the KBS is equal to that of the integer-programming model. This means that we reach the optimum solution.
The KBS has the advantage of distributing the surplus workdays evenly among days and labor.

In this work, we presented a knowledge-based system built using Visual Prolog V5.2 to solve the tour-scheduling problem in continuous-operating facilities which include hierarchical workforce.

The proposed method is very efficient since its results are the same as integer-programming models.