Resilience in Complex Systems: How Systems Survive, Adapt, and Transform

Last Updated June 6, 2026

Resilience and adaptive systems theory examines how complex systems maintain functionality, reorganize, learn, and evolve in the face of disturbance, uncertainty, stress, and structural change. Traditional stability models often assumed that systems fluctuate around equilibrium and return to balance after disruption. Resilience thinking challenges that assumption. Many real systems do not simply return. They absorb, adapt, degrade, reorganize, transform, or shift into new regimes.

In systems modeling, resilience is not merely toughness, resistance, or short-term recovery. It is a dynamic property of a system’s structure: the capacity to absorb disturbance while preserving essential functions, relationships, identity, and adaptive potential. A resilient system may recover quickly after a shock, but it may also reorganize into a new configuration that preserves core function under changed conditions.

Adaptive systems deepen this idea. They do not merely respond passively to external stress. They learn, adjust rules, alter connections, redistribute resources, change behavior, and sometimes transform their own structure. Ecosystems, infrastructure networks, supply chains, public institutions, markets, cities, health systems, and organizations all contain adaptive processes. Their resilience depends not only on redundancy or strength, but also on diversity, feedback, modularity, learning, governance, memory, and the ability to reorganize before collapse becomes unavoidable.

The modern resilience framework emerged from ecological research associated with C. S. Holling and later expanded into social-ecological systems, sustainability science, infrastructure planning, disaster risk, complex adaptive systems, and systems modeling. For formal modelers, resilience shifts attention from ordinary performance to disturbance response, recovery time, threshold proximity, regime persistence, adaptive capacity, transformability, and long-run viability.

This article examines resilience and adaptive systems as core concepts in systems modeling. It covers equilibrium and resilience, adaptive capacity, self-organization, regime shifts, adaptive cycles, panarchy, socio-technical resilience, infrastructure resilience, ecological resilience, institutional resilience, recovery time, basin stability, transformability, modeling workflows, mathematical foundations, R and Python examples, responsible use, common pitfalls, and authoritative references.

Why Resilience Matters

Resilience matters because complex systems are rarely governed by average conditions alone. They are tested by disturbance, volatility, uncertainty, compound stress, delayed recovery, and changing environments. A system that performs efficiently under normal conditions may fail under shock. A system that appears stable may be brittle. A system that resists every disturbance may become rigid and unable to adapt. A system that transforms may preserve deeper function even while changing structure.

Systems modeling uses resilience concepts to ask questions that ordinary performance metrics often miss. How much disturbance can the system absorb? How quickly does it recover? Which functions must be preserved? What feedback loops support recovery? Which thresholds would move the system into a less desirable regime? Where does redundancy improve resilience, and where does it create cost or lock-in? When should the goal be resistance, recovery, adaptation, or transformation?

Resilience is especially important in domains where disruption is not exceptional but expected. Climate stress, biodiversity loss, aging infrastructure, supply-chain shocks, cyber disruption, public health crises, institutional distrust, financial instability, and geopolitical volatility all require systems that can operate under changing and uncertain conditions.

Resilience is therefore not only a property of systems. It is also a lens for deciding what kind of system performance matters under uncertainty.

Back to top ↑

From Stability to Resilience

Early systems models often emphasized equilibrium. Under this perspective, disturbance temporarily displaces a system from a stable state, and internal feedback mechanisms return it to balance. This logic remains useful in some settings, especially where systems are well-bounded, feedback is strong, and disturbances are modest. But many ecological, institutional, infrastructural, and socio-technical systems do not behave as simple equilibrium-return systems.

Ecological research in the 1970s challenged the narrow equilibrium view by showing that ecosystems may contain multiple stable states and that disturbance may trigger reorganization rather than simple recovery. A lake may shift from clear to turbid. A fishery may collapse into a low-productivity state. A forest may recover after disturbance, or it may shift into grassland or shrubland. In these cases, stability around a single equilibrium is not enough to explain system behavior.

The resilience perspective reframed stability as only one possible feature of system behavior. A system can be stable but brittle. It can be variable but resilient. It can resist small shocks while being vulnerable to structural transition. It can preserve function by changing form. It can also remain resilient in ways that are harmful, such as persistent poverty traps, institutional corruption, ecological degradation, or infrastructure dependency.

The shift from stability to resilience is a shift from asking whether a system returns to asking how it persists, adapts, reorganizes, or transforms under pressure.

Back to top ↑

What Is Resilience in Systems Modeling?

In systems modeling, resilience refers to the capacity of a system to absorb disturbance, preserve essential function, recover from disruption, adapt to changing conditions, and avoid undesirable regime shifts. It is not a single metric. It is a family of related properties that depend on the system boundary, disturbance type, functional priority, time horizon, and scale of analysis.

A hospital can be resilient to ordinary demand variation but not to a pandemic surge. A supply chain can be resilient to local supplier failure but vulnerable to global shipping disruption. A city can be resilient to short-term power outages but vulnerable to long-term heat stress. An ecosystem can be resilient to seasonal variability but vulnerable to nutrient loading, drought, or habitat fragmentation.

This means resilience must always be specified carefully: resilience of what, to what, for whom, over what time horizon, and at what scale. A model that does not answer these questions risks turning resilience into a vague positive label rather than an analytical concept.

Resilience modeling begins with these boundary questions because resilience is not neutral. A system can be resilient in ways that preserve harm, or transformative in ways that create more just and viable futures.

Back to top ↑

Adaptive Systems and Self-Organization

Adaptive systems are systems whose components change behavior in response to conditions, feedback, learning, incentives, stress, and interaction. They do not simply follow fixed mechanical rules. They update. Households adapt to prices and risk. Firms adapt to competition. Agencies adapt to rules and constraints. Ecosystems adapt through species interactions and selection pressures. Infrastructure systems adapt through operator decisions, emergency response, maintenance, and redesign.

Self-organization occurs when system-level patterns emerge from distributed local interactions rather than central command. This can strengthen resilience by allowing flexible, local response. It can also create fragility when local adaptation produces harmful collective outcomes. For example, individual drivers rerouting around congestion may overload secondary roads. Firms protecting themselves from supply risk may create upstream volatility. Households responding to scarcity may create hoarding or price spikes.

Systems modeling can represent adaptive systems through agent-based models, system dynamics with adaptive parameters, network models with rewiring, evolutionary models, reinforcement learning structures, and hybrid simulations. The modeling task is to represent not only the system state but also the rules by which the system changes its own response.

Adaptive capacity is not automatically beneficial. Systems can adapt into brittle, extractive, inequitable, or degraded regimes. Good resilience modeling therefore asks what kind of adaptation is occurring, who benefits, and whether adaptation preserves or undermines long-term viability.

Back to top ↑

Resilience modeling is strongest when it treats these dimensions as interacting design choices rather than as a single score.

Back to top ↑

Disturbance, Exposure, and Sensitivity

Resilience cannot be evaluated without specifying disturbance. A system may be resilient to one disturbance and vulnerable to another. The same supply chain may handle small demand volatility but fail under port closure. The same watershed may absorb ordinary rainfall but not repeated drought. The same organization may handle workload variation but collapse under simultaneous staff loss and public scrutiny.

Three concepts are especially important: exposure, sensitivity, and adaptive response. Exposure describes the magnitude and frequency of disturbance. Sensitivity describes how strongly the system responds to disturbance. Adaptive response describes how the system adjusts after disturbance. A resilient system may reduce exposure, reduce sensitivity, improve recovery, or change structure to avoid repeated harm.

Good resilience models test multiple disturbances, not only the disturbance that is easiest to quantify. They also test compound events, because systems often fail when stressors interact.

Back to top ↑

Regime Shifts and System Transformation

A regime shift occurs when a system reorganizes around a new set of feedback relationships, constraints, and behavioral patterns. This is one of the central links between resilience theory and systems modeling. Resilience is not only about recovery after disturbance; it is also about whether disturbance pushes the system into a new regime.

Regime shifts can be ecological, infrastructural, social, financial, technological, or institutional. A lake may shift from clear to turbid. A city may shift from transit-oriented growth to car-dependent sprawl. A public institution may shift from credibility to mistrust. A maintenance system may shift from preventive care to emergency repair. A market may shift from liquidity to panic. A health system may shift from service delivery to overload triage.

Regime shifts are difficult because interventions that work in one regime may fail in another. Once feedback structures change, previous assumptions about response, recovery, and control may no longer apply.

Resilience modeling should therefore test not only whether the system recovers after disturbance, but whether it remains inside a desirable regime.

Back to top ↑

Adaptive Cycles and System Evolution

Resilience scholars often describe system evolution through adaptive cycles. These cycles do not imply mechanical inevitability, but they provide a useful way to think about how systems grow, become efficient, accumulate rigidity, experience disruption, and reorganize. The cycle is often described through four phases: growth, conservation, release, and reorganization.

During growth, systems expand, accumulate resources, form connections, and explore possibilities. During conservation, systems become more efficient, connected, and structured. This can improve performance, but it can also increase rigidity. During release, disturbance disrupts the established structure. During reorganization, new arrangements become possible.

The adaptive-cycle framework is valuable because it shows how success can create fragility. Systems that optimize too tightly around current conditions may lose diversity, slack, modularity, and learning capacity. When disturbance arrives, they may be efficient but brittle.

Systems modeling can represent adaptive cycles through changing feedback strength, resource stocks, network connectivity, diversity, rigidity, disturbance frequency, and reorganization rules.

Back to top ↑

Panarchy and Cross-Scale Resilience

Panarchy extends adaptive-cycle thinking across scale. Systems are nested: households within neighborhoods, neighborhoods within cities, cities within regions, regions within economies, economies within ecological systems, and institutions within historical contexts. Resilience at one scale can depend on processes at another scale.

Cross-scale interaction is crucial. A local system may appear resilient because it transfers risk elsewhere. A regional system may be resilient because local systems absorb shocks. A slow-moving ecological or infrastructural process may constrain fast-moving economic or political choices. A crisis at a small scale may trigger transformation at a larger scale. Conversely, rigidity at a larger scale may suppress adaptive change at smaller scales.

Panarchy reminds modelers that resilience cannot be understood at one scale alone. The system that appears resilient locally may be fragile globally, and the system that appears disruptive locally may support long-term transformation at a larger scale.

Back to top ↑

Resilience in Socio-Technical Systems

Socio-technical systems combine technology, infrastructure, organizations, institutions, rules, incentives, knowledge, labor, and human behavior. Power grids, hospitals, public transportation systems, emergency management networks, digital platforms, water systems, food systems, supply chains, and financial platforms are all socio-technical systems.

Resilience in these systems cannot be reduced to technical redundancy. A backup generator does not create resilience if fuel supply, maintenance, staffing, communication, and governance fail. A digital platform may have server redundancy but weak accountability. A transit system may have spare vehicles but no operator capacity. A hospital may have beds but insufficient nurses. A city may have adaptation plans but weak implementation authority.

Systems modeling must therefore integrate technical structure with organizational behavior and institutional response. Network models can represent dependency and propagation. System dynamics can represent capacity, backlog, fatigue, and recovery. Agent-based models can represent decentralized adaptation. Hybrid models can connect all three.

Socio-technical resilience depends on the alignment of infrastructure, people, institutions, and information under stress.

Back to top ↑

Network Resilience and Cascading Risk

Many resilient systems are networks. Their behavior depends on nodes, links, connectivity, centrality, redundancy, modularity, and dependency. Network resilience asks how a system continues functioning when nodes fail, links break, flows reroute, or shocks propagate.

Highly connected networks can be efficient, but they can also transmit failure quickly. Modular networks can contain disruption, but they may reduce coordination. Redundant networks can reroute flows, but redundancy may be expensive or unevenly distributed. Centralized networks can be easy to coordinate, but they may be vulnerable to hub failure. Distributed networks can be robust to local failure, but they may struggle with coordination and standardization.

Network resilience is therefore not a simple matter of more connection. It depends on the pattern of connection, the type of shock, and the system’s ability to reroute, isolate, repair, and learn.

Back to top ↑

Resilience design requires tradeoffs. A perfectly optimized system may be fragile. A highly redundant system may be expensive. A decentralized system may adapt locally but coordinate poorly. A centralized system may coordinate well but fail catastrophically at the center. Modeling helps make these tradeoffs visible.

Back to top ↑

Resilience in an Era of Global Change

Resilience thinking has become central because many systems are experiencing disturbance faster than institutions can adapt. Climate change, biodiversity loss, urbanization, infrastructure aging, migration, economic volatility, geopolitical instability, pandemics, artificial intelligence, and digital dependency all create conditions where past stability is a poor guide to future viability.

Global change is not one disturbance. It is a changing disturbance environment. Systems may face repeated shocks, chronic stress, interacting hazards, and shifting baselines. A city may face heat, flooding, housing stress, grid strain, and fiscal pressure at the same time. A food system may face climate stress, water scarcity, supply-chain disruption, labor instability, and geopolitical risk. A public institution may face demand growth, mistrust, resource constraints, and information disorder together.

Systems modeling supports resilience planning by testing how systems behave under plausible stress combinations, long time horizons, delayed consequences, and adaptive responses. It helps shift planning from “what is most likely?” to “what must continue functioning, what could fail, and how can the system adapt?”

Resilience in an era of global change is not the ability to preserve everything. It is the ability to preserve what matters, adapt what must change, and transform what cannot remain viable.

Back to top ↑

Resilience, Adaptation, and Transformability

Resilience, adaptation, and transformability are related but distinct. Resilience emphasizes persistence of essential function under disturbance. Adaptation emphasizes adjustment within an existing system. Transformability emphasizes the capacity to create a fundamentally different system when the current one is no longer viable.

This distinction matters because not every system should be preserved. A system can be resilient while maintaining inequality, ecological damage, institutional exclusion, or technological lock-in. In those cases, resilience at one scale may be harmful at another. The goal may not be to make the existing system more durable, but to enable transition toward a different regime.

Systems modeling helps clarify whether the goal should be persistence, adaptation, or transformation. It also helps reveal who benefits from each option.

Back to top ↑

Measuring Resilience

Measuring resilience is difficult because resilience is contextual. A single index can be useful for communication, but it can also hide the specific functions, disturbances, time horizons, and tradeoffs that matter. A good resilience assessment usually combines multiple indicators.

Common measures include performance loss, recovery time, area under a performance curve, probability of threshold crossing, time in degraded state, redundancy, diversity, modularity, adaptive capacity, social vulnerability, governance responsiveness, and learning capacity. In dynamic models, resilience can also be evaluated by simulating shocks and measuring how the system responds.

Resilience metrics should be interpreted as evidence for judgment, not as substitutes for judgment. The model still needs values, boundaries, and accountability.

Back to top ↑

Mathematical Lens: Recovery, Basin Stability, and Adaptive Response

A simple resilience-oriented system can be represented as a nonlinear dynamical process:

Near a stable equilibrium \(x^*\), local recovery can be approximated by:

Recovery time after a disturbance can be represented as:

A basin-stability view asks whether a disturbed state remains inside the attraction domain of the desired regime:

A simple resilience-loss process can be represented as:

Adaptive systems may change their own governing parameters:

A performance-based resilience metric can be written as:

These equations show why resilience modeling is more than return-to-equilibrium analysis. It includes recovery speed, disturbance absorption, threshold boundaries, adaptive change, and performance over time.

Back to top ↑

Back to top ↑

Strengths and Limitations

Resilience thinking strengthens systems modeling by moving analysis beyond ordinary performance and equilibrium return. It highlights disturbance, recovery, thresholds, adaptation, transformation, and cross-scale interaction. It also helps decision-makers understand why maximum efficiency can create fragility, why redundancy may be valuable, and why recovery is not always a simple reversal of damage.

At the same time, resilience can become vague if it is not specified carefully. Analysts must define the system, disturbance, function, scale, values, and beneficiaries. Resilience can also be politically ambiguous. A harmful system can be resilient. An unjust institution can persist. A fragile community may be asked to become resilient instead of receiving structural support. Responsible modeling must therefore connect resilience to purpose, equity, and accountability.

The value of resilience modeling is not that it gives a single answer. It helps decision-makers see what must remain functional, what could fail, what might adapt, and what may need to transform.

Back to top ↑

R Workflow: Simulating Resilience Loss and Recovery Time

The R workflow below uses base R. It simulates repeated shocks, weakening stability, adaptive recovery, recovery-time diagnostics, and resilience performance over time.

# resilience_adaptive_systems_diagnostics.R
# Base R workflow:
# simulating resilience loss, recovery time, adaptive capacity, and repeated shocks.
#
# Suggested repository placement:
# articles/resilience-and-adaptive-systems/r/resilience_adaptive_systems_diagnostics.R

args <- commandArgs(trailingOnly = FALSE)
file_arg <- grep("^--file=", args, value = TRUE)

if (length(file_arg) > 0) {
  script_path <- normalizePath(sub("^--file=", "", file_arg[1]), mustWork = TRUE)
  article_root <- normalizePath(file.path(dirname(script_path), ".."), mustWork = TRUE)
} else {
  article_root <- normalizePath(getwd(), mustWork = TRUE)
}

tables_dir <- file.path(article_root, "outputs", "tables")
figures_dir <- file.path(article_root, "outputs", "figures")

dir.create(tables_dir, recursive = TRUE, showWarnings = FALSE)
dir.create(figures_dir, recursive = TRUE, showWarnings = FALSE)

set.seed(42)

simulate_resilience <- function(
  scenario,
  n_steps = 180,
  initial_recovery_strength = 0.22,
  recovery_erosion = 0.0009,
  learning_gain = 0.0007,
  shock_multiplier = 1.0,
  adaptation_floor = 0.03
) {
  time <- seq_len(n_steps)
  state <- numeric(n_steps)
  adaptive_capacity <- numeric(n_steps)
  shock <- numeric(n_steps)
  performance <- numeric(n_steps)

  adaptive_capacity[1] <- initial_recovery_strength
  state[1] <- 0
  performance[1] <- 1

  shock_times <- c(25, 55, 90, 125, 155)
  shock_values <- c(1.5, 1.7, 2.0, 2.2, 2.5) * shock_multiplier
  shock[shock_times] <- shock_values

  for (t in 2:n_steps) {
    adaptive_capacity[t] <- max(
      adaptation_floor,
      adaptive_capacity[t - 1] -
        recovery_erosion +
        learning_gain * max(0, 1 - abs(state[t - 1]))
    )

    state[t] <- state[t - 1] -
      adaptive_capacity[t] * state[t - 1] +
      shock[t] +
      rnorm(1, 0, 0.025)

    performance[t] <- max(0, 1 - abs(state[t]) / 4)
  }

  data.frame(
    scenario = scenario,
    time = time,
    state = state,
    adaptive_capacity = adaptive_capacity,
    shock = shock,
    performance = performance,
    performance_loss = 1 - performance
  )
}

recovery_time_after_shock <- function(df, shock_time, tolerance = 0.25) {
  after <- df[df$time >= shock_time, ]

  recovered <- after$time[abs(after$state) <= tolerance]

  if (length(recovered) == 0) {
    return(NA_real_)
  }

  min(recovered) - shock_time
}

runs <- rbind(
  simulate_resilience(
    scenario = "baseline_adaptation",
    initial_recovery_strength = 0.22,
    recovery_erosion = 0.0009,
    learning_gain = 0.0007,
    shock_multiplier = 1.0
  ),
  simulate_resilience(
    scenario = "weakened_capacity",
    initial_recovery_strength = 0.16,
    recovery_erosion = 0.0014,
    learning_gain = 0.0003,
    shock_multiplier = 1.0
  ),
  simulate_resilience(
    scenario = "compound_stress",
    initial_recovery_strength = 0.18,
    recovery_erosion = 0.0012,
    learning_gain = 0.0004,
    shock_multiplier = 1.35
  ),
  simulate_resilience(
    scenario = "learning_investment",
    initial_recovery_strength = 0.24,
    recovery_erosion = 0.0006,
    learning_gain = 0.0012,
    shock_multiplier = 1.0
  )
)

shock_times <- c(25, 55, 90, 125, 155)
summary_rows <- data.frame()

for (scenario_name in unique(runs$scenario)) {
  subset_data <- runs[runs$scenario == scenario_name, ]

  recovery_times <- sapply(
    shock_times,
    function(shock_time) recovery_time_after_shock(subset_data, shock_time)
  )

  summary_rows <- rbind(
    summary_rows,
    data.frame(
      scenario = scenario_name,
      final_state = subset_data$state[nrow(subset_data)],
      maximum_abs_state = max(abs(subset_data$state)),
      minimum_performance = min(subset_data$performance),
      mean_performance = mean(subset_data$performance),
      final_adaptive_capacity = subset_data$adaptive_capacity[nrow(subset_data)],
      average_recovery_time = mean(recovery_times, na.rm = TRUE),
      unrecovered_shocks = sum(is.na(recovery_times)),
      cumulative_performance_loss = sum(subset_data$performance_loss)
    )
  )
}

write.csv(
  runs,
  file.path(tables_dir, "r_resilience_adaptive_system_trajectories.csv"),
  row.names = FALSE
)

write.csv(
  summary_rows,
  file.path(tables_dir, "r_resilience_adaptive_system_summary.csv"),
  row.names = FALSE
)

png(file.path(figures_dir, "r_resilience_adaptive_systems.png"), width = 1200, height = 700)
plot(
  NULL,
  xlim = range(runs$time),
  ylim = range(runs$performance),
  xlab = "Time",
  ylab = "Performance",
  main = "Resilience Performance Under Repeated Shocks"
)

for (scenario_name in unique(runs$scenario)) {
  subset_data <- runs[runs$scenario == scenario_name, ]
  lines(subset_data$time, subset_data$performance, lwd = 2)
}

legend(
  "bottomright",
  legend = unique(runs$scenario),
  lwd = 2,
  bty = "n",
  cex = 0.75
)
grid()
dev.off()

print(summary_rows)
cat("R resilience and adaptive systems diagnostics complete.\n")

This workflow shows how repeated shocks, weakening recovery strength, compound stress, and learning investment alter resilience performance over time.

Back to top ↑

Python Workflow: Modeling Adaptive Capacity Under Repeated Shocks

The Python workflow below uses only the standard library. It simulates repeated shocks, adaptive capacity, recovery dynamics, performance loss, recovery-time diagnostics, and scenario comparison.

#!/usr/bin/env python3
"""
Resilience and adaptive systems workflow.

Dependency-light workflow demonstrating:

1. Repeated disturbance shocks
2. Adaptive capacity dynamics
3. Recovery after disruption
4. Performance loss over time
5. Scenario comparison
6. Resilience diagnostics

All data are synthetic.
"""

from __future__ import annotations

from pathlib import Path
import csv
from statistics import mean
import random


ARTICLE_ROOT = Path(__file__).resolve().parents[1]
TABLES = ARTICLE_ROOT / "outputs" / "tables"


def write_csv(path: Path, rows: list[dict[str, object]]) -> None:
    path.parent.mkdir(parents=True, exist_ok=True)
    if not rows:
        raise ValueError(f"No rows to write: {path}")

    with path.open("w", newline="", encoding="utf-8") as handle:
        writer = csv.DictWriter(handle, fieldnames=list(rows[0].keys()))
        writer.writeheader()
        writer.writerows(rows)


def simulate_resilience(
    scenario: str,
    steps: int,
    initial_adaptive_capacity: float,
    recovery_erosion: float,
    learning_gain: float,
    shock_multiplier: float,
    adaptation_floor: float = 0.03,
) -> list[dict[str, object]]:
    random.seed(42)

    shock_schedule = {
        25: 1.5 * shock_multiplier,
        55: 1.7 * shock_multiplier,
        90: 2.0 * shock_multiplier,
        125: 2.2 * shock_multiplier,
        155: 2.5 * shock_multiplier,
    }

    state = 0.0
    adaptive_capacity = initial_adaptive_capacity

    rows: list[dict[str, object]] = []

    for time in range(1, steps + 1):
        shock = shock_schedule.get(time, 0.0)

        if time > 1:
            adaptive_capacity = max(
                adaptation_floor,
                adaptive_capacity
                - recovery_erosion
                + learning_gain * max(0.0, 1.0 - abs(state)),
            )

            state = (
                state
                - adaptive_capacity * state
                + shock
                + random.gauss(0.0, 0.025)
            )

        performance = max(0.0, 1.0 - abs(state) / 4.0)

        rows.append({
            "scenario": scenario,
            "time": time,
            "state": round(state, 6),
            "adaptive_capacity": round(adaptive_capacity, 6),
            "shock": round(shock, 6),
            "performance": round(performance, 6),
            "performance_loss": round(1.0 - performance, 6),
        })

    return rows


def recovery_time_after_shock(
    rows: list[dict[str, object]],
    shock_time: int,
    tolerance: float = 0.25,
) -> int | str:
    for row in rows:
        if int(row["time"]) >= shock_time and abs(float(row["state"])) <= tolerance:
            return int(row["time"]) - shock_time

    return ""


def summarize(rows: list[dict[str, object]]) -> list[dict[str, object]]:
    summary_rows: list[dict[str, object]] = []
    shock_times = [25, 55, 90, 125, 155]

    for scenario in sorted(set(str(row["scenario"]) for row in rows)):
        subset = [row for row in rows if row["scenario"] == scenario]
        states = [float(row["state"]) for row in subset]
        performances = [float(row["performance"]) for row in subset]
        losses = [float(row["performance_loss"]) for row in subset]
        capacities = [float(row["adaptive_capacity"]) for row in subset]

        recovery_times = [
            recovery_time_after_shock(subset, shock_time)
            for shock_time in shock_times
        ]
        numeric_recovery_times = [
            float(value)
            for value in recovery_times
            if value != ""
        ]

        summary_rows.append({
            "scenario": scenario,
            "final_state": round(states[-1], 6),
            "maximum_abs_state": round(max(abs(value) for value in states), 6),
            "minimum_performance": round(min(performances), 6),
            "mean_performance": round(mean(performances), 6),
            "final_adaptive_capacity": round(capacities[-1], 6),
            "average_recovery_time": round(mean(numeric_recovery_times), 6)
            if numeric_recovery_times
            else "",
            "unrecovered_shocks": sum(1 for value in recovery_times if value == ""),
            "cumulative_performance_loss": round(sum(losses), 6),
            "diagnostic_label": (
                "resilience weakening"
                if capacities[-1] < capacities[0] and min(performances) < 0.5
                else "adaptive recovery"
                if capacities[-1] >= capacities[0]
                else "managed disturbance"
            ),
        })

    return summary_rows


def main() -> None:
    scenarios = [
        {
            "scenario": "baseline_adaptation",
            "steps": 180,
            "initial_adaptive_capacity": 0.22,
            "recovery_erosion": 0.0009,
            "learning_gain": 0.0007,
            "shock_multiplier": 1.0,
        },
        {
            "scenario": "weakened_capacity",
            "steps": 180,
            "initial_adaptive_capacity": 0.16,
            "recovery_erosion": 0.0014,
            "learning_gain": 0.0003,
            "shock_multiplier": 1.0,
        },
        {
            "scenario": "compound_stress",
            "steps": 180,
            "initial_adaptive_capacity": 0.18,
            "recovery_erosion": 0.0012,
            "learning_gain": 0.0004,
            "shock_multiplier": 1.35,
        },
        {
            "scenario": "learning_investment",
            "steps": 180,
            "initial_adaptive_capacity": 0.24,
            "recovery_erosion": 0.0006,
            "learning_gain": 0.0012,
            "shock_multiplier": 1.0,
        },
    ]

    all_rows: list[dict[str, object]] = []

    for scenario in scenarios:
        all_rows.extend(simulate_resilience(**scenario))

    summary_rows = summarize(all_rows)

    validation_rows: list[dict[str, object]] = []

    for row in summary_rows:
        for metric, low, high in [
            ("maximum_abs_state", 0.0, 1000000.0),
            ("minimum_performance", 0.0, 1.0),
            ("mean_performance", 0.0, 1.0),
            ("final_adaptive_capacity", 0.0, 100.0),
            ("unrecovered_shocks", 0.0, 1000.0),
            ("cumulative_performance_loss", 0.0, 1000000.0),
        ]:
            value = float(row[metric])
            validation_rows.append({
                "scenario": row["scenario"],
                "metric": metric,
                "value": round(value, 6),
                "target_low": low,
                "target_high": high,
                "passed": low <= value <= high,
            })

    write_csv(TABLES / "python_resilience_adaptive_system_trajectories.csv", all_rows)
    write_csv(TABLES / "python_resilience_adaptive_system_summary.csv", summary_rows)
    write_csv(TABLES / "python_resilience_adaptive_system_validation_checks.csv", validation_rows)

    print("Resilience and adaptive systems workflow complete.")
    print(TABLES / "python_resilience_adaptive_system_summary.csv")


if __name__ == "__main__":
    main()

This workflow demonstrates how adaptive capacity, learning, shock magnitude, and recovery erosion can be represented as dynamic system properties rather than static labels.

Back to top ↑

GitHub Repository

Back to top ↑

Ethics and Responsible Use

Resilience is ethically complex. It can support better planning, preparedness, and adaptation. It can also be misused to shift responsibility onto vulnerable communities, normalize crisis, or preserve harmful systems. A resilience model should never imply that people should simply become more resilient to avoidable harm.

Responsible resilience modeling asks who benefits from resilience, who bears the burden of adaptation, what systems are being preserved, and whether transformation would be more just than persistence. It also asks whether resilience at one scale creates vulnerability at another scale.

Resilience should be used to expand responsibility, not to excuse neglect. The goal is not simply to make systems endure stress, but to build systems worth sustaining.

Back to top ↑

Common Pitfalls

Resilience modeling can fail when resilience is treated as a slogan, a single metric, or an inherently positive property. It can also fail when modelers ignore scale, distribution, thresholds, adaptation, or transformation.

The central correction is to treat resilience as a structured modeling question, not a generic virtue.

Back to top ↑

Conclusion

Resilience and adaptive systems theory changed systems science by shifting attention from equilibrium alone to persistence, recovery, reorganization, adaptation, and transformation under disturbance. It showed that many systems do not simply return to balance after disruption. They may absorb shocks, reorganize, cross thresholds, shift regimes, or transform into new configurations.

For systems modeling, this shift is profound. The most important question is often not whether a system is stable in ordinary conditions, but whether it can continue functioning when conditions change. A resilient model must represent disturbance, exposure, sensitivity, feedback, recovery, adaptive capacity, thresholds, learning, cross-scale interaction, and values.

Resilience thinking also adds humility. It reminds modelers that efficiency is not the same as viability, that recovery is not always reversal, that adaptation is not always good, and that transformation may sometimes be necessary. It also forces attention to distribution: resilient for whom, against what, and at whose expense?

Used responsibly, resilience and adaptive systems modeling helps researchers and decision-makers prepare for uncertainty without pretending to control it. It supports systems that can absorb, learn, reorganize, and transform while preserving the functions and relationships that make them worth sustaining.

Back to top ↑

Related Articles

• What Is Systems Modeling?
• Systems Thinking vs Systems Modeling
• Core Principles of Systems Modeling
• System Dynamics Modeling
• Agent-Based Modeling
• Network Models
• Nonlinearity, Thresholds, and Regime Change
• Critical Transitions and Tipping Points in Complex Systems
• Early Warning Signals of System Collapse
• Panarchy and Multi-Scale Systems Modeling
• Infrastructure Systems Modeling
• Environmental Systems Modeling

Further Reading

• Stockholm Resilience Centre. Research on resilience, planetary boundaries, social-ecological systems, sustainability science, and global change. Available at: https://www.stockholmresilience.org/.
• Resilience Alliance. Research network on resilience in social-ecological systems. Available at: https://www.resalliance.org/.
• Folke, C. (2006) ‘Resilience: The emergence of a perspective for social–ecological systems analyses’, Global Environmental Change, 16(3), pp. 253–267. Available at: https://www.sciencedirect.com/science/article/pii/S0959378006000379.
• Holling, C.S. (1973) ‘Resilience and stability of ecological systems’, Annual Review of Ecology and Systematics, 4, pp. 1–23. Stable record available at: https://www.jstor.org/stable/2096802.
• Walker, B., Holling, C.S., Carpenter, S.R. and Kinzig, A. (2004) ‘Resilience, adaptability and transformability in social–ecological systems’, Ecology and Society, 9(2), art. 5. Available at: https://www.ecologyandsociety.org/vol9/iss2/art5/.
• Scheffer, M. (2009) Critical Transitions in Nature and Society. Princeton, NJ: Princeton University Press. Publisher page available at: https://press.princeton.edu/books/hardcover/9780691122045/critical-transitions-in-nature-and-society.
• Dakos, V., Carpenter, S.R., van Nes, E.H. and Scheffer, M. (2015) ‘Resilience indicators: prospects and limitations for early warnings of regime shifts’, Philosophical Transactions of the Royal Society B, 370(1659). Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC4247400/.
• Gunderson, L.H. and Holling, C.S. (eds.) (2002) Panarchy: Understanding Transformations in Human and Natural Systems. Washington, DC: Island Press.
• Meadows, D.H. (2008) Thinking in Systems: A Primer. White River Junction, VT: Chelsea Green.
• Sterman, J.D. (2000) Business Dynamics: Systems Thinking and Modeling for a Complex World. Boston: Irwin/McGraw-Hill.
• System Dynamics Society. System Dynamics Society. Available at: https://systemdynamics.org/.
• MIT Sloan System Dynamics. About Us. Available at: https://mitsloan.mit.edu/faculty/academic-groups/system-dynamics/about-us.

Back to top ↑

References

• Dakos, V., Carpenter, S.R., van Nes, E.H. and Scheffer, M. (2015) ‘Resilience indicators: prospects and limitations for early warnings of regime shifts’, Philosophical Transactions of the Royal Society B, 370(1659). Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC4247400/.
• Folke, C. (2006) ‘Resilience: The emergence of a perspective for social–ecological systems analyses’, Global Environmental Change, 16(3), pp. 253–267. Available at: https://www.sciencedirect.com/science/article/pii/S0959378006000379.
• Gunderson, L.H. and Holling, C.S. (eds.) (2002) Panarchy: Understanding Transformations in Human and Natural Systems. Washington, DC: Island Press.
• Holling, C.S. (1973) ‘Resilience and stability of ecological systems’, Annual Review of Ecology and Systematics, 4, pp. 1–23. Stable record available at: https://www.jstor.org/stable/2096802.
• Meadows, D.H. (2008) Thinking in Systems: A Primer. White River Junction, VT: Chelsea Green.
• MIT Sloan System Dynamics. (n.d.) About Us. Available at: https://mitsloan.mit.edu/faculty/academic-groups/system-dynamics/about-us.
• Resilience Alliance. (n.d.) Home. Available at: https://www.resalliance.org/.
• Scheffer, M. (2009) Critical Transitions in Nature and Society. Princeton, NJ: Princeton University Press. Publisher page available at: https://press.princeton.edu/books/hardcover/9780691122045/critical-transitions-in-nature-and-society.
• Sterman, J.D. (2000) Business Dynamics: Systems Thinking and Modeling for a Complex World. Boston: Irwin/McGraw-Hill.
• Stockholm Resilience Centre. (n.d.) Home. Available at: https://www.stockholmresilience.org/.
• System Dynamics Society. (n.d.) System Dynamics Society. Available at: https://systemdynamics.org/.
• Walker, B., Holling, C.S., Carpenter, S.R. and Kinzig, A. (2004) ‘Resilience, adaptability and transformability in social–ecological systems’, Ecology and Society, 9(2), art. 5. Available at: https://www.ecologyandsociety.org/vol9/iss2/art5/.

Back to top ↑