Herbert A. Simon: The New Science of Management Decision. Harper & Brothers Publishers, New Yorkä 1960. 50 s. $ 2.50.

I denne lille populate bog, der bestar af nogle forelassninger holdt pa handelshojskoleafdelingen af New Yorks universit i foraret 1960, blaender Simon op for de mere principielle problemer, han gennem en arraekke har arbejdet med omkring muligheden for ved hjselp af elektronregnemaskiner at lede en virksomhed bedre, end det har vaeret muligt gennem det, han kalder det gammeldags system. Simon's centrale ide er, at man kan lasre en elektronregnemaskine at traffc nye beslutninger ved at efterligne den menneskelige mekanisme, der fungerer under bsning af et nyt problem. Eller med hans egne ord:

»I can now, in summary, state the central hypothesis of the theory of problemsolving I am proposing: In solving problems, human thinking is governed by programs that organize myriads of simple information processes - or symbol manipulating processes if you like - into orderly, complex sequences that are responsiveto and adaptive to the task environmentand the clues that are extractedfrom that environment as the sequences unfold. Since programs of the same kind can be written for computers, these programs can be used to describe and simulate human thinking. In doing so, we are not asserting that there is any resemblance between the neutology of the human and the hardware of the computer.They are grossly different. However,at

ever,atthe level of detail represented by elementay information processes, programscan be written to describe human manipulation, and these programs can be used to induce a computer to simulate the human process.«

Et sporgsmal, der rejser sig i denne stille en model, der pa realistisk vis afbilder den menneskelige evne til at lese pioblemer, en model som Simon kalder »a general problem solver*. Simon's svar pa dette spargsmal lyder:

»The General Problem Solver mirrors this process as follows. Its programs enable it to formulate and attack three kinds of goals:

1. Transform goals: Change a into b.

2. Reduce difference goals: Eliminate
or reduce the difference between
a and b.

3. Apply operator goals: Apply the
program (or operator or method) Q
to the situation a.

og videre siger Simon:

»Let me summarize: GPS is a program - initially inferred from the protocols of human subjects solving problems in the laboratory, and subsequently coded for computer simulation - for reasoning in terms of ends and means, in terms of goals and subgoals, about problematic situations. It is subject matter free in the sense that it is applicable to any problem that can be cast into an appropriate general form (e. g., as a problem of transforming one object into another by the application of operators). It appears to reproduce most of the processes that are observable in the behavior of the laboratory subjects and to explain the organization of those processes. On the basis of hand simulation (it has not yet run on the computer in its present general form), we can say that GPS is a substantially correct theory of the problem-solving process as it occurs under these particular laboratory conditions. tions.How general it is remains to be seen.«

Opstilling af programmer af samme type kalder Simon for heuristisk programmering. Simon gar videre med at give nogle fi eksempler, hvor man i praksis har anvendt denne form for simulation og slutter med at pege pi de udviklingslinjer han ser for fremtidig organisation af erhvervsvirksomheder:

»1. Organizations will still be constructed in three layers; an underlying system of physical production and distribution processes, a layer of programmed (and probably largely automated) decision processes for governing the routine day-to-day operation of the physical system, and a layer of nonprogrammed decision processes (carried out in a manmachine system) for monitoring the firstlevel processes, redesigning them, and changing parameter values.

2. Organizations will still be hierarchical in form. The organization will be divided into major subparts, each of these into parts, and so on, in familiar form of departmentalization. The exact bases for drawing departmental lines may change somewhat. Product divisions may become even more important than they are today, while the sharp lines of demarcation among purchasing, manufacturing, engineering, and sales are likely to fade.«

Det spørgsmål man altid møder i diskussionen af sådanne nye aspekter er, om mennesket helt kan erstattes af maskiner. Et sådant spørgsmål burde egentlig falde på sin egen urimelighed, men til folk, der alligevel stiller det, kan man sige, at man ikke kan få maskinerne til at føle og vurdere og dermed til at træffe de endelige beslutninger om, hvilket formål der bør opstilles for en organisations aktiviteter. På dette område vil mennesket stadigvæk være enerådende. Og deter iøvrigt en begrænsning, som maskinfolket må underkaste sig.