The MISG (Mathematics-in-Industry Study Group) brings together mathematicians from universities and government laboratories to work on practical problems brought from industry. Besides generating solutions to specific problems, the MISG serves to showcase to industry the problem-solving power of mathematics, and exposes academic mathematicians and students to challenges facing industry as well as providing new ideas and fresh areas of research. Since March 1993 the Department of Mathematics at the University of Melbourne has directed and sponsored the MISG.
The climax of the MISG's activities is an annual five-day workshop. On the Monday the industry representatives set the workshop in motion when they outline their projects. For the next three days the mathematicians work collaboratively in lively brainstorming sessions with the industrial presenters and on the final day the mathematicians summarise their progress. Despite the short time frame a great deal can be achieved, with participants working on the project or projects that take their fancy. MISG Summaries containing short equation-free reports, as well as detailed technical reports, are published.
MISG 1996 was held at the University of Melbourne from 29 January to 2 February 1996. A record number of delegates, over 175, attended the meeting: this number included over 50 postgraduate students and ten overseas delegates. The following eight topics were considered at the meeting.
Project 1. Warehouse optimisation (Australian Paper):
Mills which produce a range of papers of different specifications work most
efficiently in a cycle where each paper type is made every two to four weeks.
But their customers, such as the manufacturers of cardboard boxes, must be
able to respond rapidly to commercial demand. This means they may need to
switch between papers of different specifications from one
day to the next as a vegetable crop ripens, for instance, or a line of
packaging is redesigned. To overcome this mismatch in timing, a buffer of
reels of paper of different types needs to be held in a store which is
accessible to customers. While the customers maintain their
own warehouses for this purpose, these typically
cannot hold enough paper of enough different types to cope with all their
needs. So mills must use intermediate stores which increases costs, handling
and the risk of damage to paper reels. Australian Paper asked the MISG to
devise a model which would allow its paper mills to maximise direct delivery
to customer stores and minimise reel handling. The company also expected the
solution to involve recommendations as to the best way to lay out the
warehouses. In response, the MISG team was able to provide several different
approaches from which the company might be able to develop models for
delivery schedules. It was also able to recommend changes in the
traditional configuration of storage areas to increase their efficiency and
versatility.
Project 2. Methodologies for decision-making in tree breeding programs
(Various industries involved in the CRC for Temperate Hardwood Forestry):
Tree breeders have several key decisions to make
selecting which trees are to be used in the breeding
program and how often to breed them, then selecting the trees to mate with
each other and finally selecting which resultant offspring to plant out.
The MISG team was able to develop models which looked at tree breeding from
three different aspects. The first model applied to selection of trees for
breeding. It maximised the breeding value
of individual trees while minimising inbreeding. Algorithms already exist to
turn this model into a useful computer program. The second model maximised
entropy, a measure of genetic diversity. While this model needed further work,
it looked promising. The third model took a population of desired genetic
diversity and gave a measure of the breeding value of the best progeny.
After further development, it should be possible to use this third model
to predict genetic improvement from generation to generation.
Project 3. Homogenisation of photographic dispersions (Kodak):
Fine droplets of photographic chemicals are formed by forcing an
emulsion of coarse droplets through a steel tube with one or more abrupt
constrictions, each consisting of a small hole in a barrier or plate across
the cylinder. Kodak approached the MISG to examine the
optimum geometry for this device, which is known as an orifice disperser,
the best size and shape for the hole or holes, and the number and spacing
of plates. The MISG team came up with several ideas for changing the
geometry of the plate but did not have data available to test their impact.
It became clear from the team's work that, in general, the configuration
used in the disperser was relatively efficient and that there
was only a marginal gain in using more than one plate.
Project 4. Cooling of jarred cheese spreads (Kraft):
Cheese spread is a remarkably good insulator. It takes a long time to cool
after it is cooked. But if the temperature of
the spread is not brought down from about 85 degrees C to less than 25
degrees C
as soon as possible, it tends to spoil. In order to achieve the correct
cooling, Kraft Foods passes filled jars of cheese spread through a cooler
which lowers the temperature to an acceptable level within about 50 minutes.
Having recently built a new production line
for cheese spreads at its factory complex in Strathmerton, northern Victoria,
Kraft Foods asked the MISG to model and examine the operation of the cooler.
The MISG team was able to develop applicable heat transfer models
(which gave answers closely matched to Kraft's data) and recommended
changes to the cooling system including installation of a larger heat
exchanger to chill water to be used. The team also gave a nifty explanation
to why the cheese spread can boil near the centre of the jar and how
to eliminate it.
Project 5. Modelling optical fibre cable (MM Cables):
MM Cables, a major manufacturer and supplier of cable to the
telecommunications industry in Australia and overseas, asked MISG to model
the manufacture of optical fibre cable to help with setting and controlling
equipment. The MISG team derived three models of the trajectory
of the loose tubes, one estimating the position of the tube with respect to
the backbone and the other two incorporating the changes in rotational
speed of the machinery used to wind the cable. These latter two models could
be improved if more were known about the process of winding. The team
recommended extracting the necessary data from slow-motion video.
As cables are subject to tension when being laid, and may have to cope
with ambient temperatures which can vary between 70 degrees C and 9640
degrees C,
the company also asked the MISG team to investigate conditions under
which the fibres would remain strain free should the cable
expand or contract. The team developed a relatively simple model to cover
the case of expansion, but found that contraction could lead to buckling
of the fibres, making it difficult to model.
An experimental program will be established at the University of Queensland.
Project 6. Wet gum labelling of wine bottle (Southcorp Wines):
Bubbles are endemic to the labels of glass drink bottles.
They are a matter of little concern to manufacturers of soft drinks and beer.
But to Southcorp Wines, a producer of premium wines under
brand names such as Penfolds and Lindemans, bubbles in the label can ruin
the impression of quality. The problem is enigmatic. It occurs in about one
batch in 20, and appears to be dependent on moisture. Bubbles which form
under humid conditions disappear when conditions dry out. The winemakers,
label-printers and glue manufacturer were looking for an indication of
how sensitive the process was to bottle defects, the relative humidity
and water absorption of paper, and the strength of the glue. The MISG team
looked at the problem from a variety of angles and made
several suggestions. First, the team recommended better temperature and
humidity control to ensure better glue bonding. Second, the elastic nature
of the glue could cause tension as the paper expanded upon absorbing water.
So the team advised breaking up the continuity of glue to help make it more
flexible. Third, in order to set, the glue loses water. The team calculated
that the label does not absorb enough water to assist this process much.
Perhaps the bubble problem could be alleviated if the paper absorbed no
water at all, eg by using paper varnished on both sides. Fourth, the
length of time spent pressing the label onto the bottle depends on the speed
of the production line. To ensure the highly viscous glue spreads so
that labels are stuck firmly, this time must be greater than the calculated
minimum. To achieve this minimum time, the team estimated the maximum speed
of the production line.
Insert photograph here.
Mr Barry Grennan, Mr Bruce McComb, Dr Jim Banhan (Australian Paper) and Professor Alan Gilbert, Dr Kerry Landman at opening of MISG.
Project 7. Modelling the cooking process of a single cereal grain
(Uncle Tobys):
Cereal grains are seeds packed with starch granules.
Heating them in a moist environment causes the granules initially to
swell and then to burst releasing a viscous starch paste. This process is
known as gelatinisation. The starch paste is what gives
the breakfast cereal its texture. If kept at the cooking temperature,
however, the paste begins to degrade almost immediately. So, for optimum
taste, the grains should be cooked to the point of gelatinisation and no
further. Uncle Tobys Company asked MISG to determine how to cook grains for
the best texture and taste. The company was looking to speed up the cooking
process and ensure its grains were evenly cooked. More specifically it wanted
the MISG to establish two models of a single grain:
one predicting moisture content, and the other predicting internal temperature
in terms of grain properties and size, process temperature and pressure, and
time. The MISG team set up the requested models for several different grain
shapes, and gradually refined them. The team rapidly established that heating
a grain to cooking temperature took about 10 seconds, while wetting it took
about 30 minutes. So, by presoaking the grain before heating, it was possible
to cut current cooking times of between 60 and 90 minutes to less than half.
And the grain would be cooked more evenly.
Project 8. Blending methodologies in talc operations
(Western Mining Corp.):
Magnesium silicate rock which is crushed into talc
is graded according to brightness (whiteness) and content of impurities,
the most significant of which is calcium oxide. The two highest grades,
which are bright and low in calcium oxide, are in demand
and command the best price. Western Mining manually sorts talc rock into six
different grades for shipping. In the past, if there were not enough of a
specific grade in stock to fill a particular order, the company would
mine more. It now wants to take a more sophisticated approach.
The company recognises that its grading categories
are coarse and that if it could meet the specifications of orders more
precisely, it could use its materials more efficiently. Often, for
instance, the material supplied to meet an order betters the specification
requested. In this case, high grade material could be
blended with a certain amount of cheaper lower grade rock and still meet the
order requirement. Not only would this extend the high quality resource,
it would also ensure better utilisation of lower grade material.
With this in mind, Western Mining
asked MISG to build a mathematical model which would allow the company
to blend talc efficiently and determine such questions as the optimum
size for stockpiles, the best overlap of brightness and impurities for
differing products and the best ways of using the highest grade resource.
The MISG team made progress towards building
models based on two different approaches and provided ideas for a third model.
The Steering Committee for the meeting included representatives from the Departments of Mathematics at two universities, CSIRO DMS, CRA and WMC. Generous financial support in the form of travel grants was provided by CSIRO under the DMS Applied and Industrial Mathematics Program, the Department of Industry, Science and Technology under the Science and Technology Awareness Program, and four companies, Amcor, ICI Plastics, Kodak and Sola International. ANZIAM also provided financial support. DEC in Melbourne kindly loaned three alpha workstations for the week and these were used intensively on Kodak, Kraft and Uncle Tobys projects. The Science Multimedia Teaching Unit at the University of Melbourne generously sponsored the development of the MISG Web site. Please have a browse at
In January 1996, the ARC Strategic Review Of Mathematical Sciences Research and Advanced Mathematical Services in Australia released its report Mathematical Sciences: Adding to Australia . It details the state of mathematical sciences in Australia and also identifies 10-15 year goals for the discipline. Dr Noel Barton was the Executive Officer and editor of the Review and on the Tuesday he outlined the recommendations and led a lively discussion on its findings.
As usual, useful publicity for the mathematics profession was generated by the MISG. The recent highlight was ABC TV Quantum report (Nov 1) on the Kodak project brought to MISG 1995. The publicity helps community awareness of the skills and resources provided by mathematicians.
The MISG plays an important role in the Australian industrial and academic environment and helps to foster valuable and long-term links. The MISG continues to build on its success and the Department of Mathematics at the University of Melbourne is pleased to sponsor this forum. The next MISG will be held 27-31 January, 1997 at the University of Melbourne, the week before the ANZIAM meeting in Lorne, Victoria. I hope to see you there!
Department of Mathematics
The University of Melbourne
misg@maths.mu.oz.au