ix The Sciences of The Artificial was last revised, in 1981, it
is time to ask what changes in our understanding of the world call for changes in the text.
Of particular relevance is the recent vigorous eruption of interest in complexity and complex
systems.
p.35-36 Because the consequences of many actions extend well into
the future, correct prediction is essential for objectively rational choice... In simple cases uncertainty arising
from exogenous [originating outside of the local environment] events can be handled by estimating the probabilities of these
events... but at a cost in computational complexity and information gathering... A system can generally be steered more accurately
if it uses feedforward, based on prediction of the future, in combination with feedback, to correct the errors of the past.
p.55 Problem solving is often described as a search through a vast maze
of possibilities, a maze that describes the environment. Successful problem solving involves searching the maze selectively
and reducing it to manageable proportions.
p.125-126 Manheim's procedure [for solving highway location problems] incorporates
two main notions: first, the idea of specifying a design progressively from the level of very general plans down to determining
the actual construction; second, the idea of attaching values to plans at the higher levels as a basis for deciding which
plans to pursue to levels of greater specificity... Manheim's scheme for deciding which alternatives to pursue from
one level to the next is based on assigning costs to each of the design activities and estimating highway costs for each of
the higher-level plans. The highway cost associated with a plan is a prediction of what the cost would be for the actual route
if that plan were particularized through subsequent design activity. In other words, it is a measure of how "promising" a
plan is. Those plans are then pursued to completion that look most promising after the prospective design costs have been
offset against them.
In the particular method that Manheim describes, the "promise" of
a plan is represented by a probability distribution of outcomes that would ensue if it were pursued to completion. The distribution
must be estimated by the engineer... In the highway location procedure the evaluation of higher-level plans performs two functions.
First, it answers the question, "Where shall I search next?" Second, it answers the question, "When shall I stop the
search and accept a solution as satisfactory?" Thus it is both a steering mechanism for the search and a satisficing criterion
for terminating the search. [JLJ - note that a direct application to computer chess can be attempted here, based
on an evaluation function which estimates how "promising" a position is]
p.126 Let us generalize the notion for guiding search activity... The [computer]
program [playing a game] begins to search along possible paths, storing in memory a 'tree' of the paths it has explored.
Attached to the end of each branch - each partial path - is a number that is supposed to express the 'value' of the path.
But the term "value" is really a misnomer. A partial path is
not a solution of the problem... Hence it is more useful to think of the values as estimates of the gain to be expected
from further search along the path than to think of them as "values" in any more direct sense. For example, it may
be desirable to attach a relatively high value to a partial exploration that may lead to a very good solution but with a low
probability. If the prospect fades on further exploration, only the cost of the search has been lost. The
disappointing outcome need not be accepted, but an alternative path may be taken instead. Thus the scheme for attaching
values to partial paths may be quite different from the evaluation function for proposed complete solutions. [footnote
7 - That this point is not obvious can be seen from the fact that most chess-playing programs have used similar or
identical evaluation procedures both to guide search and to evaluate the positions reached at the ends of paths.]
p.127 Thus the search process may be viewed... more generally as
processes for gathering information about problem structure that will ultimately be valuable in discovering a problem solution.
p.143, 144 A second example illustrates the importance, in choosing a representation
for a design problem, of identifying correctly the limiting resource or resources... The real design problem [in the
example being discussed] is not to provide more information to people but to allocate the time they have available for
receiving information so that they will get only the information that is most important and relevant to the decisions they
will make. The task is not to design information-distribution systems but intelligent information-filtering systems.
[JLJ - Google seems to have been taking this advice.]
p.147 Good predictions have two requisites that are often
hard to come by. First, they require either a theoretical understanding of the phenomena to be predicted,
as a basis for the prediction model, or phenomena that are sufficiently regular that they can simply be extrapolated...
The second requisite for prediction is having reliable data about the initial conditions
p.147 Since the consequences of design lie in the future, it would seem that forecasting is an unavoidable
part of every design process.
p.148 Design for distant futures would be wholly impossible if remote
events had to be envisioned in detail. What makes such design even conceivable is that we need to know or guess about the
future only enough to guide the commitments we must make today.
p.149 Two complementary mechanisms for dealing with changes in the external environment are often
far more effective than prediction: homeostatic mechanisms that make the system relatively insensitive to the environment
and retrospective feedback adjustment to the environment's variation... a stock of inventories permits a factory
to operate without concern for very short-run fluctuations in product orders... A modest excess of capacity in electric generating
plants avoids the need for precise estimation of peak loads.
p.149 Feedback mechanisms, on the other hand, by continually responding to discrepancies
between a system's actual and desired states, adapt it to long-range fluctuations in the environment without forecasting.
In whatever directions the environment changes, the feedback adjustment tracks it, with of course some delay.
In domains where some reasonable degree of prediction is possible, a system's adaptation
to its environment can usually be improved by combining predictive control with homeostatic and feedback methods.
p.160 Defining what is meant by progress... is not easy. Increasing success in meeting basic...
needs... is one kind of definition that most people would agree upon. [JLJ note to self - put in paper]
p.161 From a pragmatic standpoint we are concerned with the future because securing a satisfactory future
may require actions in the present. Any interest in the future that goes beyond this call for present action has to be charged
to pure curiosity. It belongs to our recreational rather than our working day.
p.161 In addition to the things we can do to produce immediate consequences, we must anticipate the time
lags involved in developing new capital plant and the even greater time lags involved in developing the body of technology
and other knowledge that we will need in the more distant future. Attention of the decision-making bodies has to be allocated
correspondingly.
p.161 It is a commonplace organizational phenomenon that attending to the needs of the moment - putting
out fires - takes precedence over attending to the needs for [the middle-range and long range decisions].
p.162, 163 A paradoxical, but perhaps realistic, view of design
goals is that their function is to motivate activity which in turn will generate new goals... The idea of final goals
is inconsistent with our limited ability to foretell or determine the future. The real result of our actions is to
establish initial conditions for the next succeeding stage of action.
p.210 The distinction between the world as sensed and the world
as acted upon defines the basic condition for the survival of adaptive organisms. The organism must develop correlations
between goals in the sensed world and actions in the world of process. When they are made conscious and verbalized, these
correlations correspond to what we usually call means-end analysis. Given a desired state of affairs and and existing state
of affairs, the task of an adaptive organism is to find the difference between these two states and then find the correlating
process that will erase the difference.
Thus problem solving requires continual translation between the state
and process descriptions of the same complex reality.
p.211 There is now a growing body of evidence that the activity called human
problem solving is basically a form of means-end analysis that aims at discovering a process description of the path that
leads to a desired goal. The general paradigm is: Given a blueprint, to find the corresponding recipe.
