Strategic transport planning what O. R. can offer
Page 71
If you've noticed an error in this article please click here to report it so we can fix it.
IN the previous article I outlined some typical problems to which Operational Research scientists have made •valuable contributions. I believe that once transport managers get the "feel" of the kind of problems where O.R. can point the way to a solution their advice will be sought much more frequently.
Professor Alan Mercer. of Lancaster University, told the .Windermere course attended about a problem he was currently working on for the Turkish Government. It concerned cement production and distribution, the problem of the Government being to encourage cement production and use in the arid eastern half of the country. To transport cement from western Turkey production plants was expensive. A complicating factor in the East was the prevalence of a black market in cement! About half of the production of cement in Turkey was nationalized.
It was, said the professor, a problem 'tingeing on the cost of transport and distribution. He had to consider the spectrum between centralization and de-centralization, and the cost advantages of specialized or standard lorries. As with so many such problems, if it were possible to start from scratch in an entirely undeveloped country, a rational solution would more easily be contrived.
, The Turkish problem involved the siting of cement silos in the eastern part of the country. It has been found that cement clinker—transportable in bulk—could be stored in the open air adjacent to the silos. If the decision had been to transport cement in bags to the East a higher transport cost would have been incurred and local production near to the point of use would not have been encouraged. It was therefore decided to transport the clinker from western factories to grinding mills, some of which could be located at silo sites, with subsequent deliveries by road—in so far as roads existed! Communications in eastern Turkey, and, indeed between the two halves of the country, one gathered, are rather primitive.
Prof. Mercer's job was to advise on the location of new silos and determine the most economical sources of supply for the cement clinker. The vast number of alternative solutions, all with varying cost penalties, and associated practical and political problems, made the assignment a challenging one. It was easy to see why such a logistical problem was given to an O.R. scientist, and not to a transport organization.
Market statistics
Transposing some elements of the Turkish problem to Britain, Prof. Mercer described the approach of an O.R. man. Market statistics in the UK would be studied, where available, to determine the areas of maximum potential sales. How many pre-mix concrete plants could serve such areas? Where would they be sited for most economical distribution? How many mixer trucks would be needed?
By placing a rectangular grid over a map, perhaps using 30-mile squares, or using the 2000 administrative areas in the UK as reference points, it would be possible to "play with the map and find units about which you could say something positive as to their sales potential".
Such an approach was feasible for most widely distributed products. Map references were helpful in pin-pointing areas and for computer processing.
Prof. Mercer had some pertinent things to say about current preoccupations with distribution costs. Calculating the profitability of each individual customer's business could lead to ludicrous situations. For example, if a manufacturer decided that customer number 20 was being served unprofitably and should be told that no more orders would be accepted, the other 19 customers would each have to bear some additional cost. "Number 19 could be removed, on the same argument, and if the process were not soon halted you'd soon shut up shop." Some organizations behaved in this way, he insisted.
There were some rueful smiles when the professor told the students that in his investigations he frequently questioned the accuracy of the transport costs he was presented with by client firms. "Very few firms are really positive about their transport costs. By questioning the costs I often win a Few weeks relief!"
Prof. Mercer is highly knowledgeable about the inter-relationship of marketing /transport factors. What is the effect on sales of consumer products if the number of a manufacturers' branded lines stocked by a retailer is reduced? In what circumstances and to what extent can standards of
distribution service--from receipt of' order to shop delivery, or frequency of delivery—be deliberately reduced? When does it pay to raise standards? What is the value of an analysis of missed sales, due to inadeAuate retail display or inefficient distribution? "It's a great benefit to know more about your customers than about yourself," opined the professor.
Service standards may be defined in many different ways but it seems likely that own-account operators will increasingly work to the standards found acceptable in particular trades, reviewing the standards quite frequently as cost patterns change. A 90 per cent service level could mean that deliveries to customers are within X days from receipt of order at the factory, nine times out of 10. O.R. scientists like Prof. Mercer can calculate investment levels as a function of service level and product range.
Operating buses
Another speaker at the Windermere O.R. course was Mr J. C. Wilkinson, who has spent much time investigating the practicability of computer scheduling of buses and crews. In a paper presented in 1969 at a Research Conference "Operating buses", Mr Wilkinson posed some fundamental questions having regard to the continued decline, at a steady four or Eve per cent a year, of bus use.
He felt that most operating studies sought the answer to the wrong questions, ie where, how often and when ought buses to run? Which buses ought to cover which journeys? And which crews should be in which buses? The key question in Mr Wilkinson's mind was "Why ought buses to run?" Most managers of bus services, private as well as public, had ill-defined objectives he claimed.
Although Mr Wilkinson, in work for the Tilling Association, wrote a computer program which would have produced good schedules for a rural or inter-urban mixed transport fleet of 1200 buses in about two hours, this was never fully tried out. However, the program produced acceptable schedules for the 36 buses from one depot in less than three minutes.
Rather surprisingly, Mr Wilkinson is rather pessimistic about the possibility of saving money from computerized crew scheduling, though he notes in his paper a number of current attempts to achieve