Ecological Silo Guide

The Ecological Silo manages environmental dynamics that shape evolution: resources, carrying capacity, diseases, catastrophes, and environmental cycles. It simulates the selective pressures of nature.

Overview

The Ecological Silo creates a dynamic environment that prevents populations from merely optimizing for a static fitness landscape. Instead, populations must adapt to:

• Resource scarcity and abundance - Boom and bust cycles
• Carrying capacity pressure - Population size limits
• Disease outbreaks - Pathogen spread and immunity evolution
• Catastrophic events - Mass extinctions and recovery
• Environmental cycles - Seasonal and periodic changes
• Niche competition - Specialization pressure

Why Ecological Pressure?

Without environmental pressure, evolution finds solutions that work only in idealized conditions. Real-world deployment faces:

The Ecological Silo ensures evolved solutions can handle real-world variability.

Architecture

The Ecological Silo uses TWEANN controllers at three levels:

Sensors (14 inputs)

The L0 controller receives 14 environmental measurements:

Actuators (10 outputs)

The controller adjusts 10 environmental parameters:

Ecological Dynamics

Resource Dynamics

Resources regenerate and decay according to environmental parameters:

resource_level(t+1) = resource_level(t)
                    * (1 + regeneration_multiplier * cycle_modifier)
                    * (1 - decay_rate * population_density)

When resources are scarce:

• Competition intensifies (→ Social Silo)
• Stress levels rise (→ triggers epigenetic marks)
• Innovation pressure increases (→ Cultural Silo)

Carrying Capacity

The carrying capacity enforces population size limits:

%% Fitness penalty when over capacity
penalty = overflow_penalty * (1 - exp(-(density - 1.0) / capacity_softness))

Disease System

Diseases emerge, spread, and drive immunity evolution:

Spread probability:

P(infection) = density * transmissibility * (1 - immunity)

Disease status progression:

1. Emerging - Patient zero infected
2. Epidemic - Rapid spread (>5% infected)
3. Endemic - Stable prevalence
4. Declining - Immunity rising
5. Eradicated - No active cases

Immunity acquisition:

• Recovery grants +30% immunity
• Immunity decays slowly over generations
• Creates immune diversity under selection

Catastrophe System

Catastrophic events create punctuated equilibrium:

Risk calculation:

risk = base_frequency * (1 + generations_since_last/100) * (1 + stress_level)

Catastrophe severity is capped by catastrophe_severity_cap to prevent total extinction.

Recovery period:

recovery_generations ≈ severity * 50

Environmental Cycles

Sinusoidal cycles create seasonal variation:

resource_modifier = 1.0 + amplitude * sin(2π * phase)

Cycle phases:

• Spring (0.0-0.25): Rising resources, growth opportunity
• Summer (0.25-0.5): Peak resources, population boom
• Autumn (0.5-0.75): Declining resources, competition rises
• Winter (0.75-1.0): Scarcity, stress, selection pressure

Niche Competition

Multiple niches create specialization pressure:

Symbiosis

Inter-individual relationships emerge:

Symbiotic relationships affect fitness calculations and can spread through the population.

Integration with Other Silos

Outgoing Signals

Incoming Signals

Signal Examples

%% Receiving social conflict
handle_cross_silo_signal({conflict_level, Level}, State) ->
    %% High conflict depletes resources
    ResourceMultiplier = 1.0 - (Level * 0.2),
    UpdatedEnv = State#state.environment#environment{
        resource_level = State#state.environment#environment.resource_level * ResourceMultiplier
    },
    {ok, State#state{environment = UpdatedEnv}};

%% Sending scarcity signal
maybe_emit_scarcity_signal(#environment{resource_level = Level} = Env)
  when Level < 0.3 ->
    lc_cross_silo:send_signal(resource_scarcity, #{
        level => 1.0 - (Level / 0.3),
        trend => calculate_resource_trend(Env)
    });
maybe_emit_scarcity_signal(_) -> ok.

Events Emitted

Practical Examples

Example 1: Resource-Scarce Environment

Configure a harsh environment with limited resources:

Config = #{
    resource_regeneration_multiplier => 0.7,  % Slow regeneration
    resource_decay_rate => 0.15,              % Fast consumption
    capacity_softness => 0.2,                 % Hard capacity limit
    overflow_penalty => 0.3,                  % Severe overcrowding penalty
    catastrophe_frequency => 0.02,            % More frequent disasters
    cycle_amplitude => 0.5                    % Strong seasonal variation
}.

Expected outcomes:

• Smaller, more efficient networks
• Competition-driven innovation
• Robust stress tolerance

Example 2: Disease Pressure

Configure high disease pressure:

Config = #{
    disease_virulence_cap => 0.9,             % Deadly diseases possible
    transmission_rate_modifier => 1.5,        % Fast spread
    catastrophe_frequency => 0.005            % Lower base catastrophe
}.

Expected outcomes:

• Immune diversity in population
• Avoidance behaviors evolve
• Recovery speed becomes selected trait

Example 3: Stable Environment

For baseline comparison without ecological pressure:

Config = #{
    resource_regeneration_multiplier => 1.5,
    resource_decay_rate => 0.02,
    capacity_softness => 0.9,
    overflow_penalty => 0.02,
    disease_virulence_cap => 0.3,
    catastrophe_frequency => 0.001,
    cycle_amplitude => 0.1
}.

Note: Stable environments produce solutions that may not generalize well.

Tuning Guide

Training Velocity vs. Robustness

Common Issues

Recommended Starting Point

%% Balanced ecological pressure
DefaultConfig = #{
    resource_regeneration_multiplier => 1.0,
    resource_decay_rate => 0.05,
    capacity_softness => 0.3,
    overflow_penalty => 0.1,
    disease_virulence_cap => 0.8,
    transmission_rate_modifier => 1.0,
    catastrophe_frequency => 0.01,
    catastrophe_severity_cap => 0.7,
    cycle_amplitude => 0.3,
    stress_threshold => 0.6
}.

Control Loop

The Ecological Silo executes per generation:

1. Advance environmental cycle - Update phase, calculate modifiers
2. Process resource regeneration/decay - Update resource levels
3. Check catastrophe risk - Probability increases over time
4. Execute catastrophe (if triggered) - Apply mortality, resource impact
5. Spread diseases - Process each active disease
6. Check disease emergence - New pathogens may appear
7. Enforce carrying capacity - Apply overcrowding deaths
8. Trigger stress marks - Epigenetic effects if threshold exceeded
9. Emit ecological events - Notify other systems

Configuration Reference

Full Configuration Record

-record(ecological_config, {
    %% Resource dynamics
    resource_regeneration_multiplier = 1.0 :: float(),
    resource_decay_rate = 0.05 :: float(),
    resource_capacity = 1.0 :: float(),

    %% Carrying capacity
    carrying_capacity = 1000 :: non_neg_integer(),
    capacity_softness = 0.3 :: float(),
    overflow_penalty = 0.1 :: float(),

    %% Disease
    disease_virulence_cap = 0.8 :: float(),
    transmission_rate_modifier = 1.0 :: float(),
    immunity_decay_rate = 0.01 :: float(),

    %% Catastrophe
    catastrophe_frequency = 0.01 :: float(),
    catastrophe_severity_cap = 0.7 :: float(),

    %% Environment
    cycle_period = 100 :: non_neg_integer(),
    cycle_amplitude = 0.3 :: float(),
    stress_threshold = 0.6 :: float(),

    %% Symbiosis
    symbiosis_formation_rate = 0.05 :: float(),
    symbiosis_decay_rate = 0.1 :: float(),

    %% Enable/disable
    enabled = true :: boolean()
}).

API Functions

%% Start the ecological silo
ecological_silo:start_link(#{population_id => PopId})

%% Read current environment state
{ok, Environment} = ecological_silo:get_environment(PopId)

%% Manually trigger a catastrophe (for testing)
ecological_silo:trigger_catastrophe(PopId, #{type => resource_crash, severity => 0.5})

%% Introduce a disease (for testing)
ecological_silo:introduce_disease(PopId, #{virulence => 0.6, transmissibility => 0.8})

%% Get full silo state
{ok, State} = ecological_silo:get_state(PopId)

Source Code Reference

Core implementation files:

References

• PLAN_ECOLOGICAL_SILO.md - Full specification
• "Ecology: Concepts and Applications" - Molles
• "The Theory of Island Biogeography" - MacArthur & Wilson
• "Epidemiology" - Rothman

See Also

• Task Silo Guide - Population fitness management
• Social Silo Guide - Competition and cooperation
• Cultural Silo Guide - Innovation and learning
• Distribution Silo Guide - Population diversity