6.9.4 Thinking The Problem Through In Detail Using Causal Loop Diagrams And Influence Diagrams
A number of observations drawn from the diagrams produced above lead to the formulation of remedial strategies without the necessity for quantitative modelling and simulation:
- The first question that arises is: ‘to what extent are telephone orders captured in reality?’ It was found that ThreeTwoOne missed many potential orders because the telephone was not always manned. It should be intuitively obvious that to have the clerical office manned as continuously as possible is important, both to capture and action incoming orders, and to pass them to the batch plant operators who are responsible for having the orders mixed. To assist here, a relatively simple logging of calls, orders taken and time of day would be valuable in helping determine the most likely busy times. In the short term the manning problem could be alleviated by having a manager or foreman supervisor stand in when the clerk took a routine break or was busy passing orders to the batch plant.
- The second question that arises is: ‘how quickly were concrete batches being mixed?’ Delays in passing orders to the batch plant were generally short. One important exception arose when things got busy in which case orders were not passed immediately to the batch plant until the next quite period arose. This could be corrected as the workers had suggested many months earlier by the investment in a computerised order handling system. Such a system, costing $5,000-$10,000, could also be used to manage pre-bookings. Mixing time could not be varied without new plant. The availability of the batching plant (the percentage of time the plant is serviceable), could be increased by implementing a programme of routine servicing. Initially this would involve a considerable effort after hours. Though a premium would have to be paid, advice from maintenance engineers is that availability rates could be improved by 10-25%, thereby having an immediate effect on the average rate at which batches of concrete could be mixed. In the long term, this might not obviate the need to invest in new plant, but it would set a benchmark for the required performance and reliability of new plant. The availability of plant operators could be improved by hiring additional staff. However, it has been found much more cost effective to invest in training to make staff multi-skilled. Plant operators were always considered specialists. With training, it has been found that managers could also stand in here for short periods of time.
- The third question is the one which has continually occupied the minds of he directors: ‘what is the optimum truck fleet size?’ This question has been found to be rather academic, though quantitative modelling and simulation would provide valuable insights here. It was found that the real issue was to find ways of increasing the rate of delivering batches of concrete. This could be improved by reducing the time drivers were not available due to the need to take scheduled breaks. This is not a matter of making the drivers work harder: rather it is a matter of making better use of time. When drivers are on delivery tasks relatively long distances from ThreeTwoOne’s base they are frequently out of radio communications range. These happen to be the times when the drivers are most likely to need to take their scheduled breaks. A management rule had been formulated along the lines that if a driver was out of radio range, he would not take a break, rather hw would return immediately just in case there was a batch waiting for loading. Drivers frequently arrived back, having forgone their break only to find there was no batch awaiting immediate delivery. Annoyance levels rose in these circumstances, and drivers then left to take their break at a local sandwich bar. In reality, drivers frequently missed the breaks to which they were entitled. Equipping each truck and driver with a mobile telephone and requiring drivers to call the manager as soon as they completed delivery of a load greatly reduced missed breaks. During very busy times, this enabled managers to task drivers in advance of their return to base. The improvement in communications and the morale of drivers and managers alike was immediate. To fully understand the detailed dynamics of the queue structure depicted in Figure 6-14 demands that quantitative system dynamics model be built. However, many of ThreeTwoOne’s problems can be corrected without taking this step.