Much has been made of the relationship between "rocket science" and the
scheduling of sales associates in retail stores. While the term "rocket science"
is misleading, here is the short explanation starting with a historical
perspective.

In the epoch Battle of the Atlantic during World War II the good guys struggled
to convey supplies to Britain and the Soviet Union through the submarine
infested North Atlantic. A question of strategy became apparent: do you group
the transport ships together in convoy where they can be more easily protected
but are limited to the speed of the slowest ship, or do you send ships
separately? A team of mathematicians studied the problem, and concluded that the
convoy system would minimize loses. The use of convoys together with the advent
of useable radar turned the tide - but for a time the battle was a very close
call.

The branch of mathematics stemming from the wartime effort was referred to as
operations research. Broadly speaking, the purpose of O.R. is to apply
mathematics to real-world problems that are operational in nature. Typically,
the goal is to maximize or minimize something (e.g. profit, customer service,
wait time) subject to "constraints" (e.g., budget, available sales associates).

Through the mid-1970's business schools placed much emphasis on O.R., and there
were success stories. Did you know that the best mix of ingredients in animal
feed has been determined using O.R.? The petroleum industry has also made
extensive use of O.R. both to help decide where to drill for oil (and when to
stop), and to optimize refinery operations. Unfortunately, O.R. proved
cumbersome for most real world problems. The data needed to perform the
calculations was not available, and the computers of the day could not
economically carry out the extensive calculations. By the 1980's, mergers,
downsizing, and business reengineering had become the accepted method of
obtaining large cost savings - O.R. was rarely discussed.

However by the early 90's the wheel had turned full-circle. The benefits of
corporate reorganization and reengineering had been largely realized. And, for
the first time, extensive computerization in most industries meant that "input"
data (e.g., POS data) was typically available and computer power became
inexpensive and vast.

The retail store scheduling problem is ideally suited to an O.R. solution. The
objective typically is to minimize selling cost or to maximize productivity.
Constraints are in the form of budgets, availability of sales associates, shift
composition rules, and numerous other factors.

The scheduling problem is too complicated to solve by hand. This is the case
because of the large number of feasible solutions (ones that satisfy the
constraints) that would have to be manually evaluated. Supervisory personnel can
easily develop a feasible solution that looks "ok." But finding an optimal
solution or near-optimal solution by hand is almost impossible, especially when
the solution has to be frequently recomputed because of changing conditions. And
it is the difference between an "ok" solution, and a near-optimal one that can
mean the difference between profit and loss.

Fortunately, store users need not be concerned with the "rocket science" aspects
of scheduling. The user accesses QServ screens to enter employee availability
and other information; the "rocket science" calculations happen automatically
inside the system.