Wear & Heat System

How the Weibull wear model determines agent lifespan, how heat visualizes wear progress, and how the heatmap shows congestion.

The wear system models mechanical degradation in MAFIS. Each agent has a pre-sampled failure time drawn from a Weibull distribution at simulation start. As agents move, their operational age increases. When an agent’s operational age reaches its sampled failure time, it dies permanently. The heat value is a visual indicator of how close each agent is to failure.

Weibull Wear Model

At initialization, each agent is assigned a failure time using the Weibull inverse CDF:

$t_{\text{fail}} = \eta \cdot (-\ln U)^{1/\beta}$

where $U \sim \text{Uniform}(0, 1)$, $\beta$ is the shape parameter, and $\eta$ is the scale parameter.

These failure times are sampled once per agent at simulation start using the dedicated fault_rng stream. No per-tick RNG is consumed for wear detection.

Wear Presets

Death Trigger

Each tick, agents that move increment their operational_age. When operational_age >= weibull_failure_ticks[i], the agent dies:

1. Agent state → Dead
2. Agent’s cell → permanent obstacle in GridMap
3. Cascade pipeline fires (ADG → BFS → replan)
4. FaultType::Overheat event recorded

[!IMPORTANT] Wear detection is a simple deterministic comparison. There is no per-tick probability roll. Wear-based death is fully reproducible from the seed.

Heat (Visual Indicator)

Heat is not a fault trigger. It is a visual indicator showing each agent’s progress toward its sampled failure time:

$\text{heat} = \frac{\text{operational\_age}}{\text{failure\_tick}}$

• heat = 0.0 → fresh agent, far from failure
• heat = 1.0 → agent at or past its failure time

Heat is displayed as a color gradient on agents in the 3D viewport: cool colors for low heat, warm colors for agents approaching failure.

[!NOTE] Heat shows individual agent progress, not population-level Weibull CDF. Two agents with heat 0.8 are each 80% toward their own (different) failure times.

Heatmap

The heatmap is a spatial density grid updated each tick. It provides a real-time overlay in the 3D viewport showing congestion hot spots. Cells where agents frequently wait show high density. Lightly trafficked cells show low density. The heatmap has two modes: density (decaying warm gradient) and traffic (cumulative blue).

Configuration

Wear parameters live in FaultConfig (src/fault/config.rs):

pub struct FaultConfig {
    pub enabled: bool,
    pub weibull_enabled: bool,
    pub weibull_beta: f32,
    pub weibull_eta: f32,
    pub intermittent_enabled: bool,
    pub intermittent_mtbf_ticks: u64,
    pub intermittent_recovery_ticks: u32,
}

The Weibull model is activated when weibull_enabled is true. Intermittent faults (temporary latency) are controlled separately.

Code Location

• src/core/runner/mod.rs : sample_weibull_ticks(), update_agent_wear(), detect_weibull_faults()
• src/fault/config.rs : FaultConfig resource
• src/fault/heat.rs : HeatmapState resource, heatmap overlay
• src/analysis/fault_metrics.rs : fault metrics computation