GiveUpGiveInAllocationModel

Overview

The GiveUpGiveIn model is designed to be a simple mechanism for new agents to enter the model

Each timestep, the model does the following:

for each empty cell, introduces the most competitive potential agent to fill the cell (as per the SimpleAllocationModel)
numTakeovers times, the set of PotentialAgent is sampled according to their competitiveness on a perfect cell (raised to an exponent): P(a)=Comp(a)^b. This is done for every time to account for changes in supply. Then
- numCells cells are chosen at random
- they are ranked according to the competitiveness the agent would obtain on them
- the agent starts with the best cell, and attempts to force out the current agent in each one. Currently, DefaultAgents decide whether to give in by comparing the new agent's competitiveness with their own, plus their Giving In threshold. If the new value is higher, they give in.
- if it succeeds, the loop terminates

Name	Type	Default	Batch-mode ready	Descritpion
numCells	String	NaN	yes	The number of cells an agent (type) can search over to find maximum competitiveness
percentageCells	String	NaN	yes	Alternative to numCells: specify the percentage of entire cells in the region to search over.
numTakeovers	String	NaN	yes	The number of times an agent (type) can search the above no. of searches for a cell
percentageTakeOvers	String	NaN	yes	Alternative to numTakeovers: specifiy the percentage of times an agent (type) can search the above no. of searches for a cell
probabilityExponent	int	2	no	The agent's competitiveness is raised to this power

<giveUpGiveInAllocationModel numCells="10" numTakeovers="1" probabilityExponent="1"/>

* numTakeovers: how many times is an agent selected and attempts to take over a cell

numCells: how many cells an agent can search per "takeover"
probabilityExponent: the agent's competitiveness is raised to this power. pE = 0 gives equal chance, raising it makes it increasingly likely that highly competitive agents get selected over less competitive ones.