Planners have always needed to make decisions quickly and have created a vast number of rulesand techniques to support that decision making. Heuristics are decision rules or algorithms thatproduce good, workable decisions without attempting to truly optimize a model of operations.

Traditionally, heuristics were manual rules of thumb. Since I know that product A is myfastest seller and it runs the best on machine 1, I always start scheduling by putting product A onmachine 1 and figure that I will run products B and C, which run about the same on all machines, onmachine 2. Of course, sometimes I can t make enough of them on 2, and I have to take some machine 1time for B or C. Simple problem, simple logic, possibly even optimal schedules in thiscase!

With computer technology, more complex heuristics could be programmed and solutions quicklygenerated. For example, some of the best-known early finite capacity scheduling tools offered theplant scheduler a menu of heuristics to apply. One heuristic was: Improve the sequence of jobs tobe run on this machine by selecting adjacent pairs of jobs in the current candidate schedule,reversing their sequence, and seeing if overall changeover cost is reduced while still meeting alldue dates. This particular heuristic, repeated hundreds of times, is actually quite effective.Users were invited to construct compound heuristic solutions of their own by programming asequence of simple heuristics, i.e., perform this heuristic two hundred times, then perform thisstep fifty times, and so on.