saa.neighborhood

Generates neighbor solutions for simulated annealing

The EvoPER, Evolutionary Parameter Estimation for Individual-based Models is an extensible
package providing optimization driven parameter estimation methods using metaheuristics and
evolutionary computation techniques (Particle Swarm Optimization, Simulated Annealing, Ant Colony Optimization
for continuous domains, Tabu Search, Evolutionary Strategies, ...) which could be more efficient and require,
in some cases, fewer model evaluations than alternatives relying on experimental design. Currently there
are built in support for models developed with 'Repast Simphony' Agent-Based framework (<https://repast.github.io/>)
and with NetLogo (<https://www.netlogo.org/>) which are the most used frameworks for Agent-based modeling.

Antonio Prestes Garcia

evoper

Evolutionary Parameter Estimation for 'Repast Simphony' Models

Alfonso  Rodriguez-Paton

saa.neighborhood function

<dl><dt>f</dt>
<dd>An instance of ObjectiveFunction (or subclass) class ObjectiveFunction</dd>
<dt>S</dt>
<dd>The current solution to find a neighbor</dd>
<dt>d</dt>
<dd>The distance from current solution S <code>distance = (max - min) * d</code></dd>
<dt>n</dt>
<dd>The number of parameters to be perturbed</dd></dl>

Arguments

saa.neighborhood — saa.neighborhood

<dl>

<dt>f</dt>
<dd>An instance of ObjectiveFunction (or subclass) class ObjectiveFunction</dd>


<dt>S</dt>
<dd>The current solution to find a neighbor</dd>


<dt>d</dt>
<dd>The distance from current solution S <code>distance = (max - min) * d</code></dd>


<dt>n</dt>
<dd>The number of parameters to be perturbed</dd>

</dl>

saa.neighborhood: saa.neighborhood

Description

Usage

Value

Arguments