Heuristic Model and Programming Used in Decision Making | Management

In this article we will discuss about the concepts of heuristic model and heuristic programming used in managerial decision making.

Heuristic Model:

The term ‘heuristic’ means serving to discover or to stimulate investigation, in a literary sense. In its technical sense, heuristic is frequently associated with the simulation of human problem-solving techniques by means of computers.

Heuristic model refers to and identifies the situation or problem in which a good number of variables are involved and the time factor is so critical that neither the human nor the electronic system can encompass all relevant information.

To deal with this information problem, real-world decision makers resort to heuristics, or simply termed’ rules-of-thumb’. In this sense, heuristic model reduces the elements of a decision into manageable, bite-size pieces.

To put the concept logically—a rational decision is dependent on different bits of information and evaluation criteria which again require simulation of human thought processes and the ‘rules-of-thumb’ comes to play a role. Thus, it is not exactly a model but an individual’s framework of mind to a problem.

For example, an operating manager’s decision not to authorise overtime may be controlled by two of several important criteria: employees do not want overtime, and the company prefers to minimise overtime work. The use of heuristics i.e., ‘rules-of-thumb’, in this case, leads to a quick decision even though overtime work on a given day may be economically feasible and easily justified.

Heuristic Programming:

Heuristic Programming refers to the process by which computers approach human learning by simulation, resimulation of data, and probability estimation. In other words, it attempts to simulate the thinking process of a manager by storing and analysing past experience on a computer.

The heuristic approach to problem-solving is based on judgment, rather than on an exact, analytical solution. Heuristic programming searches for a satisfactory, rather than the optimal solution to a problem.

At a higher level, a manager can provide logical models and decision rules. Using this logic and decision rules, a computer programme can carry out simulations to determine the best results for action.

Heuristic programming and process may be described as follows:

1. A manager feeds MIS with his subjective probabilities.

2. These probabilities are his subjective estimates of the likelihood of a specified future event, e.g., sales over 30 crores.

3. A manager then obtains the data from events in the real world and incorporates some of them, based on his judgment, to the MIS.

4. The MIS, thereafter, revises these ‘prior’ probabilities into ‘posterior probabilities’ about the event.

In order to complete the learning process in MIS, the computer system must be provided with real-world data on events duly evaluated by a manager. Let us suppose that a manager takes an action based on a prediction of the demand for his product, only to find that sales do not meet expectations.

The computer system must be provided with:

(i) A specification of the action taken,

(ii) The prediction on which it is based, and

(iii) The performance data from the real world. The first two elements are provided by the manager and the last by the environment.

It is interesting to note that in all of the many approaches to decision analysis utilised by MIS, the creation of alternatives to be considered is left to the manager. Creativity must be developed by such process as brainstorming and other methods.

Computers may rearrange data, update probabilities, and build new models, but they are unable to creatively submit new alternatives for the decision process. Herein, comes the crucial role of the heuristic programming.

The recent trend of interest in heuristic approaches to planning has led to the development of heuristic computer programmes in the areas like:

1. Scheduling construction activity,

2. Balancing assembly lines in automotive plants,

3. Locating warehouses, and

4. Scheduling local truck delivery to many small customers, etc.

In relation to the problem of warehouse locations, heuristic programming could help in a profit-oriented solution by equating the marginal costs of operating the warehouse with the cost savings and profit increments to be realised.

The heuristic programme, in this case, involved consideration of plant-warehouse location with the relation to raw materials, the kind of warehouse needed, and even shipping and inventory, arrangements as preferred.