Key Definitions

Forward Scheduling (Due Date Setting)

A feasible due date is determined by forward scheduling from the earliest release date, i.e. the date by which an order is expected to have been confirmed and materials made available. 

The forward scheduling methodology presented here matches required and available capacity overtime. It bases on the methodology presented by Bertrand (1983a) and (1983b) and Bechte (1994) and has been chosen due to its outstanding performance compared to alternative rules suitable for high variable production environments as typical for small to medium sized make-to-order companies (see Thürer et al., 2012b). 

For the method presented here a dynamic factor to allow for the queuing time F(W,C) that is dependent on both the cumulative total workload (W) and capacity (C) until a future time bucket at the work center which performs a certain operation is added to the operation due date of the job that would otherwise be determined based on infinite capacity. The operation due date based on infinite capacity is given by the operation due date of the previous operation (ODDi-1)plus the processing time (p) of the job.  The starting operation due date is given by the earliest release date plus an estimate for the expected pool waiting  time. Following Little's Law, this expected waiting time pior to release can be indicated by the workload waiting in the pool for the work center that is more likely to restrict the release possibilities; i.e. the work center with the maximum pool load across the work center(s) in the routing of the order (see also COBACABANA).

This procedure is repeated at the next work center in the routing of the job until all operation due dates have been determined, with the last operation due date becoming the due date of the order. 

 Note that the due date determined presents the internal or production due date. This due date represents the current production capability. To account for variability during production (which will always occur) an extra time buffer has to be added to obtain the lead time to be quoted to the customer. This so-called external lead time buffer allows management to determine the service level to be achieved (see e.g. Enns, 1995; Hopp & Sturgis, 2000). 

Backward Scheduling (Due Date Feasibility) 

The feasibility of a due date specified by a customer is determined by backward scheduling to find a latest release date, i.e. the date by which the order must be released from the pre-shop pool if it is to be delivered to the customer on time. 

Under backward scheduling (assuming capacity to be finite), the required capacity at each work center is loaded into the available capacity as done for forward scheduling above except the procedure works backwards from a given due date. Three options have been presented in the literature for handling an overload (i.e. backward scheduling that results in a latest released date prior to the earliest release date): 

• adjust capacity (e.g. Hendry & Kingsman 1993; Bechte, 1994; Wiendahl, 1995); 

• reject the order (e.g. Kingsman et al., 1989; Hendry & Kingsman, 1993; Corti et al., 2006); and/or, 

• change the due date by reverting to forward (re)scheduling (e.g. Kingsman et al., 1989; Bechte, 1994).

Strike Rate Analysis

An make-to-order company’s future workload is dependent on the number of tenders or quotations that are converted into firm orders: the strike rate.  A simple two-step process for developing a matrix that links strike rates to a set of outcomes was developed in the literature (see Kingsman et al., 1993, 1996; Kingsman & Mercer, 1997): 

This matrix of order winning history is updated as more historical data becomes available, i.e. after each acceptance/rejection decision. If the company wins a tender, then the value in the cell corresponding to that set of outcomes (e.g. due date and mark-up) is increased by one. But as the company would also have likely won the tender with a more competitive bid (e.g. a lower price or earlier due date), cells with a lower mark-up and/or due date are also increased by one (see the shaded area in the Figure). The strike rate for a certain set of outcomes can then be estimated as the quotient of the value in the corresponding cells and the total number of bids or tenders. 

Forward/Backward Scheduling & Strike Rate Estimation  

The forward/backward scheduling element of CEM can be used to set or assess the feasibility of due dates and incurred production costs to ensure quotes are competitive but with realistic due dates and prices which reflect the actual operational capabilities of the firm. A set of decision criteria – e.g. the expected profit, given by the mark-up multiplied by the strike rate estimate – can then be developed using the matrix of order winning history to support the bidding process. This allows profitability to be maintained while controlling the planned workload through realistic due dates. 

From an operational point of view, this provides a control loop that: (i) reduces variability – as the incoming workload is controlled – allowing the capacity buffer to be minimized; and, (ii) regulates the ‘over-booking’ strategy of a company necessary for ensuring the utilization of scarce capacity. More specifically, forward/backward scheduling and strike rates interact as follows: 

• Forward Scheduling and Strike Rates: When the workload starts to increase, the lead times that can be realistically offered become longer – this reduces the strike rate of the company and, therefore, its workload. Then, as the workload decreases, the lead times that can be realistically offered start to shorten – this improves the strike rate and, therefore, increases the workload. 

• Backward Scheduling and Strike Rates: While most authors have looked to handle overload through capacity adjustments, order acceptance/rejection or rescheduling in isolation, the real power of CEM lies in its ability to combine the three together. In fact, few companies will reject the offer of an order from a prospective customer; instead, a longer lead time or higher price will be quoted so the company can either fit the workload to the current capacity or afford to increase capacity (e.g. through overtime) to meet the workload requirements. But by increasing the lead time or price, the likelihood of winning a particular tender is reduced. Thus, instead of explicitly making the acceptance/rejection decision internally, the company passes the decision to the customer whilst implicitly controlling it through the competitiveness of the quote. 

Thus, customer enquiry management acts as a first filter which undertakes some coarse smoothing of the planned workload. This workload smoothing or load balancing corresponds to one of the main principles of heijunka in lean operations which relates to leveling peaks and troughs in the production schedule, thereby creating some stability in the workload through balancing. Creating this stability has been argued to be the only way of realistically creating a continuous flow (e.g. Liker & Meier, 2006) and facilitates subsequent control tasks as order release and shop floor dispatching.