While Workload Control principles can be operationalized with a simple sheet of paper, it is often coupled with the use of a computer system to support Workload Control-related decision making: a Decision Support System.
An example for the operationalization of Workload Control for a shop with six work centers and a set of randomly generated orders is described below. Note that this is a simple example to show how easy workload control principles can be implemented.
Workload Control is operationalized using five Excel © sheets: (i) Customer Enquiry, which creates the order in response to a customer enquiry determining processing times, routing characteristics, operation due dates, due dates; (ii) Unconfirmed Orders, which contains all unconfirmed orders; (iii) Pool Orders and Order Release, which controls the release of orders to the shop floor; (iv) Shop Floor, which contains all orders currently on the shop floor; and (v) Order Winning History, which contains all orders (whether won or not) which passed the customer acceptance/rejection decision. Orders (or Jobs) pass sequentially from one sheet to the next. Each sheet will be explained in what follows. We added Figures which update automatically if input data is changed to support the decision process. To make workload control work, this is not required. The basis is the load calculation and the control of the workload.
The Excel © file can be found here.
decision_support_system.xlsCustomer Enquiry
When a request for quotation is received, a job is scheduled considering available capacity and current workload. The workload and capacity measures are given in cells E26:O31. The Figure above these measures summarizes the cumulative total workload (including the load of the new job) and available capacity. The cumulative total workloads are calculated from the respective sheets in cells B41:O65 and Q41:AD59. While the latter contains the individual shop floor, pool and unconfirmed workload does the former aggregate these loads into the workload measures typically applied within workload control: the total, planned and shop floor workload. The cumulative available capacity until a certain time (i.e. the total capacity times the Overall Equipment Effectiveness) is determined by the user in the grey marked cells in B41:O65.
The order itself is created in cells B36:W36. In this simple example a job’s information consists of its identity number (ID), Planned Release Date (PRD), due date, strike rate estimate and the routing and processing time information. Each order follows the standard format thus it can easily be cut and paste from workload measure to workload measure (i.e. one sheet to the other). In order to schedule an order, first determine the strike rate. Then the work center is selected in cells E18:E23. Cells E26:O31 and the figure are automatically updated being the cells linked by an If logic. The operation processing time is then inserted in one of the cells E29:O29 which represent the time buckets for the operation due dates. The workloads and figure are automatically updated. Shifting the time bucket or changing the processing time allows the cumulative workload to be fitted into the cumulative available capacity. Once the operation is scheduled, the position of the operation in the routing, the processing time and the operation due dates are inserted in the cells corresponding to the work center in cells B36:W36.
Once all operations have been scheduled, the order is cut and pasted into the Unconfirmed Order sheet until its confirmation.
Unconfirmed Orders
This sheet contains all unconfirmed orders. In addition columns AB-AG and AK-AP calculate and provide information required at Customer Enquiry. AB-AG gives the days until the operation due date for each operation and AK-AP gives the workload which contributes to the total workload, which is the original load multiplied by the strike rate. Additional measures for Customer Enquiry are calculated in column AB-AG and AK-AP.
Once an order is confirmed it is cut and pasted into the Pool Orders and Order Release sheet. Both, orders which are confirmed and not confirmed are also copied into the Order Winning History sheet and marked accordingly.
Pool Orders and Order Release
The corrected aggregate load on the shop floor, the corrected load contribution of jobs currently selected for release and the resulting workload are given in cells H17:S19. The workload norm for each work center is determined in cells H16:H19. Jobs from the pool are selected for release through column C. The Figure and load contributions in H17:S19 update automatically being linked by If logic. In addition column B can be used to create a specific priority value for jobs in the pool. Jobs can easily be sorted using the sorting function of Excel ©. Additional measures for Customer Enquiry are calculated in column AM-AR and AV-BA; the corrected aggregate load for order release which would be contributed by each job in the pool is calculated in column AD-AI.
Once order release is completed, the selected orders are cut and pasted to the Shop Floor sheet.
Shop Floor
Completed Operations are marked in cells AA-AF. Operations marked as completed are not considered when the corrected aggregate load for order release and the load for Customer Enquiry are calculated in columns AK-AP and BC-BH, respectively. The columns AT-AY give the days until the operation due date for calculations at Customer Enquiry.
Once an order is finished, it is deleted. Alternatively, additional data could be collected – e.g. on realized operation due dates – for further analysis e.g. on schedule deviations (see e.g. Soepenberg et al., 2008).
Order Winning History
This sheet summarizes the order winning history (i.e. whether a quote was accepted or rejected) and the outcomes quoted for the strike rate analysis.
Key Definitions
Decision Support System
Workload Control adoption is often coupled with that of a computer system to support Workload Control-related decision making: a Decision Support System.
The design of a Workload Control based Decision Support System should follow a structure specific design. This design approach consists of three elements: goals, means and tasks (Johannsen, 1995; Park & Lim, 1999). In other words: the definition of the typical goals of a company implementing a Decision Support System; the means of meeting these goals (i.e., through the Workload Control concept); and, the tasks or roles of humans within the domain of the Decision Support System. The tasks then define how the Decision Support System should be designed; such a task-oriented approach is considered by many to be the key to successful Decision Support System design (see, e.g., Johannsen, 1995; Johannsen, 1997; McKay & Wiers, 2003).
Defining goals starts with the following questions: What are the strategic objectives of the company? And, why does the company need a Decision Support System?
Defining the appropriate means to achieve the goals starts with the question: How can Workload Control contribute to achieve these objectives or goals?
Defining the tasks to be accomplished to effectively use the means to achieve the goals starts with the question: How is Workload Control to be used to achieve the goals?
For example one important Workload Control task is to release the right job at the right time without violating a pre-established level of workload to balance the workload. In line with authors such as Higgins (1996), McKay & Buzacott (2000) and Barthelemy et al. (2002) - who argued against a strict computer-based approach to scheduling - it is argued that, in practice, the user should be the center of the decision as to which job to release. Instead of permitting a human user to alter or intervene in computer-generated schedules, the user should actively participate in the generation of the schedules. Nonetheless, production planning tasks not only consist of elements which need special attention but also routine elements (Fransoo & Wiers, 2006). Therefore, and in order to reduce the cognitive workload of the human user, the Decision Support System should also offer the option of making the decision for the human, e.g., by automatically releasing jobs for the remaining load after the human user has released the jobs that he or she considers most important.
It has been acknowledged that Information Systems often fail to meet pre-implementation expectations (Szajna & Scamell, 1993; McKay & Buzacott, 2000; Calisir & Calisir, 2004), perhaps because most systems are designed and built without considering human factors; and, because most systems are generic tools which are not customized to company-specific needs (McKay & Buzacott, 2000; McKay & Wiers, 2003). Workload Control is a simple yet effective production planning and control solution which principles can be implemented on a sheet of paper or using a spread sheet as in the example outlined here. This makes it easy to create your own customized Decision Support System which reflects the particular needs of your enterprise.
Arguably, the important factor is that the user feels that the Decision Support System has a strong ‘usability’; in other words, that it is user-friendly, supports the accomplishment of the tasks and is efficient in achieving the goals.
User Interface
The most important criterion that should be considered for the design of the user interface can be summarized in three groups as follows:
Design and Functionality: The DSS design should support task perception and performance (Higgins, 1996; Park & Lim, 1999). The design and functionality of the system should be consistent throughout all layers of the human-machine interface (Marcus, 1992; Park & Lim, 1999). Short and diverse tasks should be avoided and the cognitive workload of the user should be kept low (Oborski, 2004).
Presentation: An appropriate level of aggregation for the information presented in the Decision Support System is required (Higgins, 1996; Oborski, 2004); for example, by using data charts or graphs. This means that the user does not have to remember data from other screens which would otherwise imply an unnecessary cognitive workload; at the same time, maintaining a clear design and avoiding small graphical objects (that prove difficult to read) is also important.
Human Factors: Support and guidance to help the user understand the system (Lin et al., 1997; Park & Lim, 1999) should be provided. Human errors should be prevented and corrected (Park & Lim, 1999; Oborski, 2004), e.g., by warning if system parameters are violated. Exits should be clearly marked and short cuts provided (Lin et al., 1997).
Recent studies have underlined the strong link between the perceived aesthetic appearance of a human-machine interface and the perceived and experienced usability (e.g., Szajna & Scamell, 1993; Tractinsky, 2000). There is also a strong link between user satisfaction and perceived and experienced aesthetics and usability (Tractinsky, 2000). Therefore, a Decision Support System should not only follow the criteria outlined in the bullet points above but also primary rules for the design of aesthetically appealing human-machine interfaces. The main primary rules are: consistency, clarity, simplicity and familiarity (Marcus, 1992). Considering the main criterion for the design outlined above, each is described in more detail below:
Consistency: Each unit of the layout grid (i.e., the grid which subdivides the computer screen) should have visual, conceptual and functional integrity. Data, functions and tools should be organised and presented appropriately to meet the criterion outlined in groups one and two above (i.e., Design & Functionality and Presentation).
Clarity: Where possible, lines of text should be kept short, legible and readable. Letters of serif type, e.g., Times New Roman, should be used where long text is necessary while short text should be presented in non-serif letter types, e.g., Arial. Colors should be used with discretion and extreme colors should be avoided. Whereas colors might be more ‘enjoyable’, they do not improve learning or comprehension more so than the use of black and white.
Simplicity: Efficient and simple navigation possibilities should be provided. Data should be presented in a simple intuitive and easy-to-understand-format. In response to the criterion outlined in group three (Human Factors) error messages should be clear and simple.
Familiarity: The user should feel familiar with the navigation possibilities and the design of the Decision Support System. Therefore it should follow standards as, e.g., Windows © or Linux ©.