cobweb2-2016-04-29

Learning

• Completely replaces the old learning agents
• Fitness function is minimizing energy losses
• Future outcomes of decisions with high energy losses are randomly modified
• Phenotypes are available to change various learning parameters

Consumption of products

• Producer uses energy to create product
• Producer can only keep N unsold products
• Consumer can pay producer for product
• Consumer will receive temporary multiplier to one of its attributes

Toxins

• Agents start with an initial amount
• Toxin amount can decrease by fraction specified per time step
• Effect on phenotypes, with multiplier and starting threshold
• Foods can contain different amounts of toxins for every agent
• Percentage transferred when eating other agents
• Percentage transferred to child
• Average population toxin level logged

Regional log file

• Logs agent and food population into a TSV file
• Numbers include cell totals for all agents and individual types
• Added option to scale bar graphs on global or local basis

Waste as food

• Agents can consume specific waste types
• Energy gained (or lost if negative) is definable
• Adds another energy event on the energy overlay

Fix loading of population

• Loading populations used to not load genetics, abiotic, and other plugin factors
• Loading works correctly now, with some limitations
• New simulation must have similar-enough global settings to original simulation
• Error will be shown when incompatible, with option to skip incompatibilities or abort