Information Technology Reference
In-Depth Information
3.4
The Satisficing Model (Nebel 2011)
Satisficing is a classic strategy in the decision-making literature ([19]) that forms an
alternative to optimizing strategies and goes back to Herbert Simon ([20], [21]). In
this concept alternatives are evaluated sequentially, in the order in which they appear
in the choice set. “Instead of hopelessly searching for the best” ([22], p. 17) a decision
maker continues seeking a solution only until he finds one that is “good enough”
([23]). The satisficing model version uses a multidimensional aspiration level as a
basis for evaluating different land-use patterns. Four thresholds (one for each of the
four objectives profit, risk, labor and legality) are defined. A solution is considered
satisfactory if and only if all thresholds are reached.
Given that this model simulates a span of 50 years, fixed underlying aspiration
levels cannot be assumed. This model hence uses flexible thresholds that are
influenced by comparison with others and former experiences of a farmer. First, a
yearly increase of the aspiration level for economic returns is implemented. Between
1960 and 2009 the gross domestic product of Spain increased almost every year ([24])
and it is assumed that farmers in the UGB want to be part of this economic growth.
Additionally Simon ([21], p. 883) states that "people tend to aspire for a future that is
little better than the present”. Based on these considerations in each step each
farmer's profit threshold is increased. Second, an adjustment of thresholds to past
outcomes is implemented. In each step and for each objective a farmers' average
value in the past five years is computed. The thresholds are then adjusted towards this
average.
The satisficing concept includes that maybe no satisfactory solution will be found.
Decision makers, who are forced to take a decision but unable to find any satisfactory
solution have two options: extending their search area or decreasing their aspiration
level. In this model both options are considered. If the current situation is not
satisfactory anymore due to changes in thresholds or external circumstances, farmers
start to search for alternative land-use patterns. If all of the alternatives which they
consider are not satisfactory, one randomly chosen threshold is decreased.
Figure 4 shows simulation results of the satisficing model that best fit
5
empirical
data (figure 2 and 3). The amount of extracted water is similar to empirical
observations with extractions just before 1990 of about 600 Mm
3
; a subsequent drop
to 200 Mm
3
followed by an increase of extractions up to 400 Mm
3
. Simulation results
show mainly three irrigated types of crops: vineyards, traditional cereals and high
yield cereals. Those crops belong to the most irrigated ones empirically observed as
well. Empirical data and model results show a similar area of irrigated melons at the
end of the simulated time span. However, the model is unable to replicate the
intermediate boom of irrigated melons and the subsequent drop around 1990.
Additionally it is not able to model the irrigation of garlic and sugarbeet correctly
which are underestimated before 1990 and overestimated towards the end of the
simulation time.
In the satisficing model the effect of laws and policies is immense. The water act
(1985) and the CAP (1993 onwards) reduce irrigation and water extraction to a large
5
The model was fitted manually. For a sensitivity analysis see [11].
