III. Basic Concepts
The planner as a decision maker is rational in Savage's sense. That is, he or she behaves according to Savage's assumptions and postulates of rationality and is a utility maximizer. In the simplified planning situation of the unitary organization, two worlds are considered: the grand world and the planner's world (a small world). The planner's world is, in Savage's language, a microcosm so that the criterion of maximizing expected utility can be applied to making choices among acts.
Each of the two worlds is described by a collection of states, actions available to the decision maker, and the consequences resulting from these actions. The states in small worlds are subsets of the states in the grand world, or the states in the grand world are the elementary states for small worlds. Let the grand world be, for example, the outcome of tossing two coins simultaneously. Note that Savage used the example of eggs conditions to explain the small world and other concepts. My using the example of tossing coins here may be more concrete. A small world would be the outcome of one of the two coins, regardless of the outcome of the other. The state of the small world, a head for example , is a subset of two states in the grand world (i.e. the outcome of the other coin can be either a head or a tail).
Events are subsets of the states in a small world. In the example of tossing a coin, the outcome of a head is an event. The outcome of either a head or a tail is also an event. Subjective probabilities are assigned to these events to express the planner's degrees of belief in the occurrence of such events. For example, the probability that a head occurs is usually 1/2. The probability that either a head or a tail occurs is one.
A decision is the planner's choice of acts (or actions) from a given set in a small world to solve problems and to anticipate desired consequences. An act will result in consequences which in turn are expressed as a function of the act. The consequences resulting from the act chosen in the small world can be realized in the grand world as acts which in turn results in consequences in the grand world. A small world could evolve resulting in transformations from one world to another. All these concepts can be expressed in functional terms as will be shown in the following sections.
A planning problem is thus a problem of modifying the current states in the planner's world to achieve the desired states, taking into account future contingencies. This modification would require a set of acts, or decisions or a plan, which would in turn result in unexpected consequences. Strictly speaking, a plan is a set of related decisions or choice alternatives conditional on events with various probability distributions. A planning activity is a set of actions taken by the planner in the planning process. I consider the case in which planning occurs in a discrete time frame in that a plan is made in each time period for solving the planning problems anticipated during that period. A new plan may or may not be formed for the next time period while the old one discarded. The process continues until the planner is satisfied with the current states or the planner's resources are depleted. A discussion of the dynamic planning problem is beyond the scope of the paper. Because of limited space I focus here instead on information processing for a single time period. That is, I study how information should be gathered in making plans for the unitary organization.