Desktop version

Home arrow Computer Science arrow Innovation Networks for Regional Development: Concepts, Case Studies, and Agent-Based Models

Internal Coherence Verification

In this section we describe simulative experiments settled to assess the internal coherence of the proposed agent-based model and we discuss simulation results. The experiments are settled to be confirmative. The calibration of the model is work in progress and will be completed in the following steps of the research.

Parameters Setting

The verification of logical internal coherence of the model is based on the assumptions derived from current body of knowledge on innovation systems framed as complex learning systems. In particular, we settled 12 different experiments (see Tables 3,4, 5, 6 and 7). In order to ensure the robustness of the results we performed

Table 3 Fixed parameters

Fixed parameters

Values

Competence (c)

Random

Noise (n)

10%

Length of message (l)

50

Acceptance-threshold

80%

Number of agents

50

Max-ticks

10,000

Runs

30

Table 4 Parameters setting to test the behaviour of a high specialized system in a dynamic environment

Parameters

I SET

II SET

III SET

Capacity of Exploration (p)

0.1

0.5

0.9

Scope (s)

0.5

Volatility (v)

0.2

Table 5 Parameters setting to test the behaviour of a high specialized system in a static environment

Parameters

IV SET

VSET

VI SET

Capacity of Exploration (p)

0.1

0.5

0.9

Scope (s)

0.5

Volatility (v)

0.8

Table 6 Parameters setting to test the behaviour of a low specialized system in a dynamic environment

Parameters

VII SET

VIII SET

IX SET

Capacity of Exploration (p)

0.1

0.5

0.9

Scope (s)

0.8

Volatility (v)

0.2

Table 7 Parameters setting to test the behaviour of a low specialized system in a static environment

Parameters

X SET

XI SET

XII SET

Capacity of Exploration (p)

0.1

0.5

0.9

Scope (s)

0.8

Volatility (v)

0.8

  • 30 runs for each experimental set. The 12 experiments are performed changing the values of the following parameters:
    • • The volatility v of the Competitive Environment CE: (v = 0.2 and 0.8)
    • • The exploration capability p of the ICAs: (p = 0.1, 0.5 and 0.9)
    • • The level of specialization s of ICAs: (s = 0.5 and 0.8)

Other parameters remain fixed according to the values of the Table 3.

Table 8 The output variables of simulations

Output variable

Description

Surviving ICAs (%)

Average (on 30 runs) number of surviving ICAs at the end of simulation as percentage of initial population

Average number of surviving Frames of ICAs

Average (on 30 runs) number of Frames in the individual memories of ICAs, calculated for each new market cycle (for each new message/Regularity provided by the CE)

Collective Interpretations’ dimension

Total number (on 30 runs) of Individual Interpretations contributing to successful Collective Interpretations

Mean Delta Budget in the system

Average (on 30 runs) value of the difference between the final budget of the system and the sum of initial budgets attributed to ICAs populating it at the beginning of simulation

 
Source
< Prev   CONTENTS   Source   Next >

Related topics