Keywords

1 Introduction

Traditional crisis management methods usually consider the crisis management system to be implicit, where the crisis framework is assumed to be the usual (or everyday) environment, with the priority given to the search for the appropriate measures to contain and then make the crisis recede as soon as possible.

This is the “Common Operational Picture” describing the situation at a specific moment in time t, which, as (Wybo and Latiers, 2006 [14]) point out, is difficult to construct, even in a crisis management center.

Beyond the different perceptions of the players concerned by the crisis (i.e., decision-makers, victims, rescue personnel, public opinion), intervention speed is almost systematically viewed as the qualitative crisis management factor.

However, with hindsight, while intervention speed is fully vindicated in terms of tactics, the move to the strategic level may require a substantial contextualizing effort. This is why SANCTUM - Crisis Anticipation by Uchronic Modeling Process - endeavors to introduce a preliminary phase in crisis analysis, which consists of making explicit the fundamentals of the human environment called into question by the crisis. This clarification – in the case of a major crisis – must go beyond the analysis of the crisis’s functional consequences on economic and social life, which is usually confined to ensuring business continuity, even under deteriorated conditions, and must not avoid ethical issues.

This clarification is required because any major disaster in one location may have domino effects in many other locations, as described in Voisard and Petrie, 2019 [12]. Each of these follow-on emergencies will be responded to by local authorities. This occurs, for example, during inter-sectoral or systemic crises, with domino effects affecting separately or simultaneously various matters of concern to society (e.g., health, energy, transportation, housing, economy, education, or public order). In France, these responses will be conducted at the territorial level such as the Operational Centers of French prefectures or local administrations. In France, the Inter-ministerial Crisis Center - in French: Cellule Internationale de Crise (CIC) - is responsible for coordinating this level of crisis management.

But how are limited national resources to be allocated? How are the territorial resources to be coordinated? It is not enough to try to pre-plan all of these possibilities, and it is not feasible to preempt local decisions about emergency tactics. Rather SANCTUM provides a common framework for working through a disaster, before and during.

The added value of SANCTUM comes to bear in times of complex crises during which severe challenges - national crises - necessarily require prioritization because of domino effects of the initial crisis and the need for coordination of responses by distributed authorities. SANCTUM addresses prioritization by providing a rational framework on which all parties can agree prior to a disaster, so as to minimize fighting for resources. A major contribution of the SANCTUM framework is to abstract what emergency responses are trying to do in all cases, and to realize that prioritization of responses must be based on cultural values. How are these values to be adjudicated during the emergency?

The SANCTUM framework provides a common process for questioning the overall meaning that the concerned human society intends to give to its existence and future and by what means of intervention and representation it can express this meaning. This allows the crisis responders to set the priorities necessitated by the scarcity of resources specific to any crisis situation and to implement the decision-making processes (Cf. (Anderson, 2010 [1]). SANCTUM also proposes an adjudication process that does not depend upon local political decisions but rather shared cultural values decided at the tactical case level by designated people in various roles. Politicians will still ultimately be in charge but their options will be well informed and transparent and the consequences of their decisions will be known by all affected parties.

Finally, the SANCTUM abstract framework of general crisis management is so well-defined as to be almost mathematical. This has two advantages. One is that ambiguity will be minimized among the participants, facilitating common understanding and cooperation. The other is that this framework can more easily be translated into various kinds of computer systems that can automate at least some of the planning and reinforce human decision-making.

This approach, which constitutes the SANCTUM project’s basic substance, offers a highly instructive analysis grid of measures taken throughout the Covid-19 crisis.

Part of this work has been published as reports of the project (Després, 2019 [3]; Fertier et al., 2019 [4]; November and Gueben-Veniere, 2019 [7]; Olié J.-L., 2019 [9]; Voisard and Petrie, 2019 [12]) or at international conferences (November et al., 2020 [7]; Petrie and Voisard, 2019 [13]), however, an overall description of the novelty of the approach has not been published so far.

This paper is organized as follows. Section 2 provides the fundamentals of the system that we propose. Section 3 defines our notion of “goal” and crisis as well as “oracles” to control the overall process. Moreover, the SANCTUM mechanism is described. Section 4 presents the concepts of progress loops and uchronias (alternative scenarios). Finally, Sect. 5 draws our conclusion.

2 The Fundamentals of the Proposed Method

The organization or agency impacted by the crisis will naturally seek to place itself in a defensive situation. It will thus adopt a defensive posture in order to mitigate the observed vulnerabilities. This posture necessarily includes constraints that are themselves sources of new vulnerabilities. In other words, the management of a crisis - initially exogenous - engenders another crisis, but this time endogenous.

Recent events provide us an example of this type of situation. Leaders responded to the pandemic (exogenous) by mandating strict lockdown measures (endogenous), which generated a diversity of concerns the dimensions of which are hard to gauge. Lacking comprehensive control of the crisis’s effects and the counter-effects of the response measures, crisis management can turn out to be, at minimum, inefficient and, at worst, it may aggravate the problem.

The response proposed by SANCTUM consists of defining an automated crisis management system, which seeks to describe as completely as possible the fundamentals of the defending organization, which are presented thereafter.

2.1 Defending Organizations

We consider the following defending organizations:

  • the system is the human, physical, legal, and technological framework in which crisis unfolds. It brings together the defending organization and its organization;

  • the values are definitions, essentially moral, which create public popular support at the time of the crisis (Schwartz, Shalom H., 2006 [10]) bear in mind that said values, by nature highly stable, are nonetheless designed by the automated crisis management system as likely to evolve; they will thus be configurable in our modelling;

  • the governance is made up of the bodies (institutions) that ensure the system’s continuity in accordance with its values (Galam, S., 2012 [5]);

  • the supporting assets are the material or intangible means that enable the system to continue functioning on an ongoing basis (e.g., economic and social infrastructures) or which support governance (e.g., public services);

  • the players are the specific or specialized human resources who develop or operate the assets; they intervene, either on the governance level or with their own strategy, which may not entail cooperation;

  • the stakes are the vital functions which are essential to ensure the system’s sustainability. The fulfilment of these functions established the conditions for the constitution of the assets and the organization of the players (e.g., housing, heating, healthcare, education). The term “impact(s)” means the total or partial loss of mentioned vital functions (or stakes) following an event in which support assets are damaged.

In the final analysis, the SANCTUM automated crisis management system brings forward a coherent interlocking of its constituents, highlighting the key notion of stake. The calling into question of the stakes is what triggers a crisis. It is not the attack on the supporting assets, the value of which fluctuating over time (for instance, in 1940, the Maginot line was a major asset, but the issue was national defense).

2.2 The Concept of Goal and a Practical Definition of the Crisis Concept

So far, we used the term “crisis” without defining it, since, in our crisis management system, this concept’s definition is not established early in the decision process – as a sort of intrinsic dysfunction that only gets worse causing organizations to become overwhelmed – but later as a systemic deduction stemming from the analysis of the impacted organization (Lagadec, P., 2008 [6]).

We have made a static description of our automated crisis management system. In reality, the system that it supports is in perpetual motion to ensure, among other things, its continuity. This self-sustaining dynamic enables it to attain the goals necessary for its development, because, as we have seen, we exclude that the defending organization remain in a vegetative or go into a regressive state.

These goals must concretely reflect the search for maximum value satisfaction. They constitute the roadmap for governance. Once the crisis breaks out, they become the stakes to be safeguarded. The stakes themselves are a function of the supporting assets and conditioned by the interaction of the players. The goals to be determined are thus a function of the time concerned (health, education and material well-being), but their intrinsic consistence must be considered constant (Massimi and Tononi, G., 2018 [11]).

The following is a case in point. If the goal is the quality of human development, one stake is education whilst the supporting assets are the locations where education takes place, and the players are those who promote or hinder education. The crisis, in the context proposed by SANCTUM, may then be seen as a disturbing phenomenon of the crisis management system that could call into question its goals.

The advantage of this definition, compared to the traditional, more qualitative ones referring more to the impairment of an equilibriumFootnote 1, is that it is somewhat measurable (the state of crisis could be the gap between the situation and the goal) therefore suitable for modelling.

2.3 Practical Examples of the Anticipation Process

Table 1 shows two simplified examples of SANCTUM’s anticipation process: one entailing probabilistic risk and the other, a terrorist threat.

Table 1. Two examples of the anticipation process
Full size table

3 Control Process of the Automated Crisis Management System

The automated crisis management system’s development must be carried out in a controlled way to ensure its convergence towards uchronia, which will become the object of a decision proposal. We must exclude the fact that there is no solution found because decision-makers cannot avoid making a decision. In this section, we present the notion of oracles as well as the five stages of the SANCTUM approach, i.e., the kernel of our methodology.

3.1 Oracles

What we refer to as the Oracles intervene on the automated crisis management system’s dynamic. We distinguish between four different types of oracles:

  • the Wise men, who control the values-linked choices. These choices require compromises (e.g., risk acceptability thresholds). They can amend the governance rule to be established, if need be;

  • the Judges, who apply the rules validated by the sages to estimate the impact and exposure levels and issue verdicts at this level;

  • the Analysts, who determine, in the most quantitative way possible, the issues, the vulnerability of assets, the search for actions and inventory of available resources;

  • the Spin doctors (communicators and influencers), who define the influence measures to be performed in the defending organization or vis-à-vis players;

The oracles, by virtue of their expertise, focus on carrying out corrective measures in their respective fields of competence in accordance with the automated crisis management system’s values and the rules set forth by the Judges.

3.2 Stages in the SANCTUM Process

Let us imagine ourselves in the anticipatory operating conditions of a crisis situation.

The first step consists of identifying our automated crisis management system, notably, how the crisis is deemed to call into question the goals of the defending organization, undermine its values and handicap its governance.

The second step, which is meant to be more concrete, consists of identifying the stakes the impairment or destruction of which is likely to call into question the automated crisis management system’s goals. The results will allow us to deduce the list of supporting assets and players concerned by said stakes.

During the third step, SANCTUM gives way to classic risk analysis.

As seen in Fig. 1, the high-stakes assets may be undermined by the vagaries of (1) probabilistic occurrence (hazards) and/or (2) determinist threats (malevolent).

Fig. 1.
figure 1

Risk generation

Full size image

In the first case (hazards), the asset’s exposure level L is:

$$ {\text{L}}\, = \,{\text{Probability}}\,*\,{\rm{Destructive}}\,{\rm{force}} $$

In the second case (malevolent threat), it is:

$$ {\text{L}}\, = \,{\text{Feasibility}}\,*\,{\rm{Attractiveness}} $$

In both cases, the impact I is the measure of harm, i.e., the damage to one or more of the system’s vital functions or stakes: loss of human life, production loss or decrease in activity, and so on.

Conventionally, risk R is expressed as follows:

$$ {\rm{R}}\, = \,{\rm{L}}\, \times \,{\rm{I}} $$

We can thus build a risk diagram on which we can locate each supporting asset according to its L and I coordinates. It is displayed on Fig. 2.

Fig. 2.
figure 2

Risk diagram

Full size image

The locus of the constant product of L × I is a hyperbola whose positioning on the plane depends on the value R-0 set to R. This determination is fundamental because, by delineating the “crisis area” on the bi-dimensional risk space, it attributes a level of severity to the crisis. The oracles attribute this value of R-0 and factor in the defending organization’s admissible level of suffering, its societal effects, the scientific and technological response capacities, etc.

We should keep in mind that, at this stage, this risk analysis is based on information “with a finite useful lifespan”, which is equivalent to the time during which the data is deemed “stable” (Wybo, 2013 [15]). The crisis management time is a structuring item of the dynamic and adaptive anticipatory approach (November et al. [7]), which must be taken into account throughout the SANCTUM process.

4 Progress Loops and Uchronias

This section presents two major concepts of our approach, namely concept loops and uchronias.

4.1 The Concepts of Progress Loops and Response Measures

SANCTUM’s “progress loop” is set to undergo digital modelling. It takes as a starting point the risk mapping the creation of which was described in Sect. 3. This mapping may be viewed as the conceptual projection of the crisis situation. In what follows, we assume that the supporting assets and players have their own temporal dynamics, the determinants of which can be known and modelled.

This modelling may initially be basic (e.g., linear changes as a function of time) or more complex (e.g., group or individual modelling of behavior to any laws of evolution). This complexity can today and, a fortiori, in the future, be taken on thanks to progress in the field of AI and possible learning from Big Data (Bénaben et al., 2008 [2]).

The next step in the SANCTUM process (Step 4) thus consists of performing risk reduction work by extracting a certain number of major supporting assets in the area at-risk and by reducing the exposure level or the impact of those that cannot be extracted.

Fig. 3.
figure 3

Definition of the priority area

Full size image

Figure 3 shows the area we should concentrate on, the priority area. It leads to the notions of mitigations and opportunities. These two types of measures consist of protecting the stakes:

  • by reducing the risks via measures on the level of the impact and exposure level of supporting assets or players;

  • by proposing new solutions when the risk-reduction measures are not successful in preventing risk materialization.

Mitigations consist of triggering a change in the system so as to reduce the level of risk, but by only intervening on factors currently known by the system. We may speak of endogenous evolution.

Opportunities assume the intervention of an outside factor, the effects of which are likely to reduce the risk level. It may be predictable, but this evolution is initially weighted by a low probability or unanticipated risk level. It may seek to adjust to the consequences of the destruction suffered to make acceptable another form of system organization.

The difference in nature between these two types of measures may be illustrated by the management of a power output crisis during exceptionally cold weather. Mitigations may be measures to protect generators so as to minimize production loss. On the other hand, higher-than-expected temperatures could constitute an opportunity to restructure the power transmission organization and eliminate weaknesses stemming from certain facilities.

4.2 Concept of Uchronia

The “progress loop” is a recursive phenomenon where the system re-assesses the overall situation and tests the crisis exit at the end of each loop.

The crisis exit is fundamentally decided by governance, but the latter intervenes in the final stages. The Sages, intervening in the early stages, assess quantitatively the convergence toward the desired goals.

Convergence is measured by comparing the changes in the indicators stemming from the adjustments (mitigations and opportunities) to the projected changes and by performing a projection of said changes over time. The system also draws from the measures taken to strengthen its self-learning and maximize the configuration of the progress loop, if the latter needs to be relaunched.

The oracles carry out these functions on the operational level:

  • the Wise men, who assess the extent of the damage suffered and the gap to be closed in order to meet the goals and the time needed to return to “peace mode”;

  • the Analysts, who use the data produced and set forth the efforts to be asked of the players or the assets.

If the model converges, it will be able to produce crisis exit scenarios. To “scenario”, we prefer the term “uchronia”, deemed to be devoid of any subjective bias.

If the model does not converge, the Wise men must recalibrate the crisis, who establish a new value R-0. The progress loop is then relaunched, with fewer resolutory constraints, which assumes a heavier weight of the automated crisis management system (level of suffering or economic or social costs on the rise).

The progress loop ends up converging and the model provides us three scenarios:

  • the worst case,

  • the likely case,

  • the optimal case.

Figure 4 shows the hyperbole of risk threshold together with the three types of scenarios mentioned above.

The worst-case scenario may be defined as the accumulation of the impacts, if all the risks in the priority zone and those selected in the priority area materialize.

The likely scenario may be defined as the accumulation of the impacts, if all the risks in the priority area and those selected in the resilience zone materialize.

The optimal scenario may be defined as the list of residual impacts, if all the risks remaining in the priority area and those remaining and selected in the resilience margin zone materialize.

Fig. 4.
figure 4

Risk mapping and representation of alternate scenarios

Full size image

4.3 Summary: The Five Steps of the SANCTUM Approach

Below are the five steps of our approach before taking a decision.

Preliminary Step

Definition of the role, responsibilities, and relationships of the anticipation cell’s participants.

Note that the mobilization of common intelligence requires maximum fluidity in exchanges. The participants must perfectly understand their place in the system and identify the roles of the other participants. Relationship quality and efficiency are key to the success of the cell’s work.

Step 1: Stakes and the Crisis Management System

Sharing of thoughts among participants about the situation.

Identification of the stakes.

Qualification of the crisis management system (i.e., values, governance, goal at stake).

Step 2: Analysis of Stakes

Further analysis of the stakes (e.g., public order, health, education, energy, transportation, or economic performance).

Identification of the assets and players concerned by these stakes.

Step 3: Risk Analysis

Conventional analysis of risks to which the assets at-stake are exposed;

Risk = level of (asset’s) exposure * impact sustained

Note that the exposure level depends on the risk type (deterministic or probabilistic).

Preparation of crises map showing the “crisis area”.

Step 4: Treatment of the Crisis Area

Preparation of measures allowing for the reduction of the crisis area by using the assets:

  • by mitigation (internal actions in the system in crisis);

  • by taking advantage of opportunities (external actions).

The implementation of mentioned measures and the projection of their effects enabling the emergence of evolving “alternate scenarios” (objectified scenarios).

Step 5: Description of the Evolving Alternate Scenarios

Understanding of the anticipated with the following three alternate scenarios:

  • the worst case (maximum impact);

  • the likely case (maximum exposure level);

  • the optimal case (maximization of the mitigation measures and opportunities).

Conclusion: Decision-Taking

The authority makes crisis-management decisions based on the description of alternate scenarios.

5 Conclusion

In this paper, we described the main concepts on which SANCTUM relies: the defending organizations, the notion of goal, oracles, as well as the stages in the SANCTUM overall process.

Even though it has been tested in real situations, SANCTUM remains a conceptual project for the time being, but one that lays the groundwork for a new crisis management approach, the need for which has become particularly apparent in the course of the Covid-19 pandemic. Its purpose is to show that, in spite of the complexity of organizations, all the functions of an automated crisis management system can be analyzed as a flow of iterative processes controlled through a “progress loop”, which benefits from a self-learning system using artificial intelligence-specific resources to assess and improve performances.

SANCTUM’s innovation consists in proposing a comprehensive breakdown of the automated crisis management system in which the crisis unfolds by considering from the outset, for example, that the values of the society affected by it are configurable and that the decision-making processes can be rendered perfectly objective by bypassing humans in favor of the Oracles.

This resolutely rational approach does not dehumanize the crisis management because, by endeavoring to remove the human from the cogs of crisis management – where s/he can be as disruptive as s/he is productive – it leaves the human the key role of final decision-makers, but with the essential predictive tools, alternate scenarios – uchronias - at their disposal.

The SANCTUM approach has been tested during the first waves of the COVID-19 crisis, as shown in the appendix.