Strategic Decision Making:

blessinghomoeopathΤεχνίτη Νοημοσύνη και Ρομποτική

30 Νοε 2013 (πριν από 3 χρόνια και 7 μήνες)

74 εμφανίσεις

Strategic Decision Making:

A Systems Dynamic Model of a Bulgarian Firm

David L. Olson, University of Nebraska

Madeline Johnson, Univ. of Houston
-
Downtown

Margaret F. Shipley, Univ. of Houston
-
Downtown

Nikola Yankov, Tsenov Academy of Economics

Transition Economies


Transition from
centrally
-
planned to
market economies


Face ambiguous
information and cues


Challenge existing
ownership & operating
principles


Firms responsible for
strategic decisions


Joint Effort


University of Houston
-
Downtown


NSF Grant


Joint
International Workshop on
the Use of Information
Technologies in Modeling
the Bulgarian Firm in
Transition from a Planned
to a Free Market Economy


Tsenov Academy of
Economics


Svishtov, Bulgaria

Subjective System Dynamics Model


Winery


Tool to simulate impact of key strategic
decisions:

1.
Market selection (local, national, international)

2.
Promotion & pricing

3.
Product quality decisions

4.
Capacity (vineyards and bottling)

5.
Distribution

Open Systems Theory


Ludwig von Bertalanffy


An organization exists in relation to its
environment


There is a continuous
flow of energy &
information


System features:


Self
-
organization

-

progressive differentiation


Equifinality



initial condition doesn’t matter


Teleology



systems are purpose
-
driven


Cybernetics


Stafford Beer


Cybernetic systems are
complex
,
probabilistic
,
self
-
regulatory
,
purposive
, have
feedback
and control


Operations research

only works when you
consider the whole


Viable System Model



organization
regulated, learns, adapts, evolves, or doesn’t
survive


Mental Models


Systems

consist of
interacting parts

working
toward
some end
,
feedback control


Purposive


Synergistic


Complex


Feedback

System Dynamics


Jay Forrester


Developed technique for deterministic
simulation of systems


Identify influences


Estimate effects


Develop feedback model

Forrester’s World Dynamics
Model


Sectors


Population


Natural Resources


Capital Investment


Pollution


Metrics


Quality of life


Material standard of living


Ratios for FOOD, CROWDING, POLLUTION

Soft Systems Theory

Peter Checkland


Interpretive action research


Model interacting system

1.
Define problem

done

2.
Express situation

done

3.
Root definition

4.
Conceptual model

done


simulation model

5.
Compare model/real world

6.
Use model to determine improved methods

7.
Action

Simulation Approaches


DYNAMO/Ithink/Stella/PowerSim


VENSIM


Commercial implementation of system dynamics


Support conceptual modeling


EXCEL


Probabilistic simulation over time


CRYSTAL BALL


Probabilistic simulation output

Development of Model


Symposium in Svishtov, Bulgaria


May 2002


About 20 from U.S., 20 from Svishtov


Selected winery because of knowledge of
Tsenov Academy faculty


Selected system dynamics because:


Problem involved subjective data


Complex interactions among decisions, time

Winery Model


Time frame: 6 years


Show impact of strategic
decisions


Inputs:


Promotion


Pricing


Quality (grow or purchase
grapes)


Market selection (local,
national, international)


Outputs


Profit


Cash flow


Market share by product (3
levels of quality)

Promotion


Lagged over three month


Impact differentials


0.5 prior month


0.35 two months prior


0.15 three months prior


Media: firm representatives interacting with
distributors


Management could constrain local, national, or
export markets to emphasize others


Demands in each market probabilistic

Quality


If winery controls vineyard, quality higher


Constrained by amount of hectares in vines


Three years between planting, use


Use own grapes as much as possible


Any extra production capacity used for purchased
grapes (lower quality bottles)

System Variables


Exogenous:


System Variables:


Control Inputs:

Exogenous Variables


Demand (normally distributed, change per month)


By market (local, national, export)


By product (correlated)


Seasonal


Market Price (normally distributed, change per month)


Independent of firm decisions


Competitor promotion (normally distributed by market)


Market share possibilities


Prior market share multiplied by ratio of prior promotion to base
promotion, divided by that of competitors


Crop yield

Control Inputs


Price


By product by month


Promotion


By product by month


Plant Capacity


Depreciation, plus construction


Labor


Permanent (higher quality) vs. temporary

System Variables


Sales


By market, by product


Inventory


High, low quality


Bank Balance


5% gain on positive balance, 15% cost on
negative

Results


Varied prices, promotion levels


Price: base, cut 10%, increase 20%


Promotion: base, emphasize local, emphasize export


Measured


bank balance after 6 years


Probability of losing initial capital (going broke)


Probability of breaking even


Market share (low, high quality)

Base Run

Wine Model
-0.2
0
0.2
0.4
0.6
0.8
1
1
4
7
10
13
16
19
22
25
28
31
34
37
40
43
46
49
52
55
58
61
64
67
70
Key parameters
index
Balance
DemN
DemEx
MktShNat
MktShEx
Base Model


1000 replications


Crystal Ball software


Cyclical demand for high quality


Base case has National focus


Without pricing & promotion, loss

End Bank Balance

Frequency Chart
lev
.000
.018
.035
.053
.070
0
1.75
3.5
5.25
7
-35,000.00
-26,250.00
-17,500.00
-8,750.00
0.00
100 Trials
100 Displayed
Forecast: endyr6 bank balance
Bank Balance


Mean 117,458 Lev


Probability of losing bankroll: 0.0


Probability of losing money: 0.0


Most optimistic:


Worst: loss:

Market Share
-

National

Frequency Chart
proporti on
.000
.015
.030
.045
.060
0
1.5
3
4.5
6
0.00
0.10
0.20
0.30
0.40
100 Trials
100 Displayed
Forecast: Market Share - National
Mixed Price, Promotion

Frequency Chart
l ev
.000
.007
.014
.020
.027
0
6.75
13.5
20.25
27
-10,000.00
-1,250.00
7,500.00
16,250.00
25,000.00
1,000 Trials
1,000 Displayed
Forecast: endy r6 bank balance
National Market Share



Mixed policies

Frequency Chart
proporti on
.000
.006
.012
.017
.023
0
5.75
11.5
17.25
23
0.30
0.36
0.43
0.49
0.55
1,000 Trials
1,000 Displayed
Forecast: Market Share National - end year 6
Model Validation


Initial visit May 2002


3 day workshop to build model


Built model summer 2002


Followup visit October 2003


Went over model in detail


Refined model structure


Identified detailed data needs

Conclusions


System dynamics useful to model subjective
input, complex interactions in temporal
environment


Need for validation