Environment is one of the most pervasive features of human existence. Every agent exists within an environment, and every action an agent takes occurs under the conditions imposed by that environment. The physical surroundings in which an agent lives, the institutions that govern them, the people with whom they interact, the technologies available to them, and the resources to which they have access all influence what the agent is able to do.
Yet despite the obvious importance of environment, it is often treated as little more than background. We speak of an agent acting within an environment, as though the environment were merely the stage upon which the agent’s decisions take place. This perspective obscures a more fundamental property of environment: it establishes conditions that constrain and enable the actions available to the agent.
A road network determines which physical movements are possible. A legal system determines which activities are permitted. An economy determines which resources can be acquired and exchanged. A social environment determines which relationships and institutions are accessible. A technological environment determines which capabilities can be exercised. In each case, the environment does more than influence the outcome of an action. It constrains the set of actions available to the agent.
This raises a more fundamental question:
What problem does an agent face by virtue of being embedded in an environment?
Answering this question first requires understanding what an environment actually is. An environment cannot simply be defined as everything that exists outside of an agent. Such a definition would make the concept so broad as to be of little analytical value. What matters is that an environment establishes the conditions under which an agent can act and pursue future states.
An environment can therefore be understood in terms of the possibilities it makes available and the constraints it imposes. Different environments expose an agent to different resources, opportunities, risks, institutions, relationships, and forms of activity. Consequently, the same agent may possess dramatically different possibilities depending on the environment in which they operate.
This creates a fundamental problem for any agent seeking to improve their position. An agent cannot simply choose any environment they desire. Access to an environment may itself require resources, transportation, knowledge, social connections, legal authorization, or other capabilities. An environment that would improve an agent’s position may therefore remain inaccessible until the conditions required to enter and operate within it have been acquired.
The problem is consequently not merely the selection of an environment. It is the acquisition, maintenance, and improvement of access to environments that increase an agent’s capacity to reach desirable future positions.
This article develops the environment problem from that perspective. It examines the relationship between agents and the environments in which they operate, the role of environmental access in determining possible futures, and the fundamental constraints involved in changing one’s environment. The objective is not to develop a strategy for any particular country, visa, or form of migration. Those are downstream problems. The objective is to understand the more fundamental problem an agent faces in acquiring and maintaining access to environments that can improve their position.
The central proposition developed in this article is:
The environment problem is the problem of acquiring, maintaining, and improving access to environments that affect the set of future positions available to an agent.
An agent does not act from an abstract or unconstrained state. At any point in time, the agent occupies some position within the game and operates under a particular set of environmental conditions. These two aspects of the agent’s state must therefore be distinguished.
Let the position of an agent at time $t$ be represented by the vector
\[P_t=(x_{1,t},x_{2,t},\ldots,x_{n,t}).\]Each $x_i$ represents some variable that describes a property of the agent’s position. The position vector does not, however, imply that every property affecting the agent must be represented as an independent component of $P_t$. Whether a particular variable constitutes a dimension of position is itself a question that must be established.
Let the environment in which the agent operates at time $t$ be represented by
\[E_t\in\mathcal{E},\]where $\mathcal{E}$ is the set of possible environments in the game.
The distinction between $P_t$ and $E_t$ is important. The position vector represents the state of the agent within the game, while $E_t$ represents the external conditions under which the agent acts. An agent may therefore possess the same internal characteristics while operating under substantially different environmental conditions.
The environment influences the actions available to the agent. Let
\[\mathcal{A}(P_t,E_t)\]denote the set of actions available to the agent at position $P_t$ while operating in environment $E_t$. In general,
\[\mathcal{A}(P_t,E_1)\neq\mathcal{A}(P_t,E_2)\]for different environments $E_1$ and $E_2$.
This distinction is fundamental. If two environments produce different sets of available actions for an otherwise identical agent, then the environment is not merely a passive backdrop to the agent’s decision problem. It is part of the conditions that determine which transitions through the game are possible.
Let the transition produced by an action $a_t$ be represented by
\[P_{t+1}=F(P_t,E_t,a_t).\]The environment therefore enters directly into the transition dynamics of the agent’s position. Two agents with identical positions $P_t$ and identical actions $a_t$ may reach different subsequent positions when operating in different environments:
\[F(P_t,E_1,a_t)\neq F(P_t,E_2,a_t).\]Likewise, an action may be available in one environment but unavailable in another:
\[a_t\in\mathcal{A}(P_t,E_1) \quad\text{while}\quad a_t\notin\mathcal{A}(P_t,E_2).\]Environment therefore affects the agent in at least two distinct ways. It can change the consequences of an action, and it can change the set of actions from which the agent can choose.
This gives us a more precise interpretation of the statement that environment constrains an agent. The environment does not necessarily determine the agent’s behavior or future position. Rather, it determines part of the conditions under which transitions from the agent’s current position are possible.
The agent’s state can consequently be represented as a sequence of position-environment pairs:
\[(P_t,E_t) \rightarrow (P_{t+1},E_{t+1}) \rightarrow (P_{t+2},E_{t+2}) \rightarrow\cdots\]where both the agent’s position and its environmental conditions may change over time.
The significance of this distinction is that the environment can change the future positions reachable from an otherwise identical position. For an environment $E$, let
\[\mathcal{R}(P_t,E)\]denote the set of positions reachable from $P_t$ under that environment. Then different environments may produce different reachable sets:
\[\mathcal{R}(P_t,E_1)\neq\mathcal{R}(P_t,E_2).\]The important question is therefore not simply which environment an agent currently occupies. It is which environments the agent can access, which transitions between environments are possible, and how those environmental transitions affect the future positions available to the agent.
The environment affects an agent not merely by determining the consequences of actions, but by constraining the actions and transitions available to the agent. For a given position $P_t$ and environment $E_t$, let
\[\mathcal{A}(P_t,E_t)\]denote the set of actions available to the agent.
The set of available actions is generally dependent on the environment. Thus, for two environments $E_1$ and $E_2$,
\[\mathcal{A}(P_t,E_1)\neq\mathcal{A}(P_t,E_2)\]may hold even when the agent’s position is otherwise identical.
This means that environment can alter the structure of the agent’s possible transitions. If an action $a$ is available in $E_1$ but unavailable in $E_2$, then the set of positions reachable by the agent differs between the two environments.
Define the set of positions reachable from $P_t$ in a single transition under environment $E$ as
\[\mathcal{R}_1(P_t,E) = \left\{ F(P_t,E,a) \mid a\in\mathcal{A}(P_t,E) \right\}.\]Two environments may therefore produce different sets of immediately reachable positions:
\[\mathcal{R}_1(P_t,E_1) \neq \mathcal{R}_1(P_t,E_2).\]This distinction is more important than simply saying that one environment is “better” or “worse” than another. An environment may make positions reachable that would otherwise be inaccessible.
The constraints imposed by an environment can take many forms. Physical geography can constrain movement. Infrastructure can constrain transportation and communication. Institutions can constrain economic activity. Laws can constrain permissible actions. Social structures can constrain relationships and opportunities. Markets can constrain access to resources. Technology can expand the actions available to an agent while its absence can restrict them.
These constraints need not be absolute. An environmental constraint may instead increase the resources required to perform an action. If an action $a$ requires a quantity of some resource $c(a,E)$, then the same action may have different costs in different environments:
\[c(a,E_1)\neq c(a,E_2).\]Thus environmental differences can affect both the feasibility and the cost of transitions.
This gives environment three fundamental effects on the agent’s decision problem:
\[\text{Environment} \rightarrow \begin{cases} \text{available actions}\\ \text{cost of actions}\\ \text{consequences of actions} \end{cases}\]The environment therefore participates in determining the set of future positions available to the agent. This is important because, within the decision framework, the utility associated with a position depends in part on the future positions that can be reached from it.
More generally, let
\[\mathcal{R}(P_t,E)\]denote the set of positions reachable from $P_t$ over the relevant future horizon under environment $E$. The single-transition reachable set is therefore a subset of the broader set of future possibilities:
\[\mathcal{R}_1(P_t,E)\subseteq\mathcal{R}(P_t,E).\]If two environments produce different reachable sets,
\[\mathcal{R}(P_t,E_1)\neq\mathcal{R}(P_t,E_2),\]then the same position $P_t$ is associated with different future possibilities under those environments.
This distinction matters because the future possibilities available from a position are part of what determines its utility. An environmental change can therefore alter the future possibilities associated with a position without requiring an immediate change in the variables represented by $P_t$.
An environmental change can be represented as
\[E_t\rightarrow E_{t+1},\]with the corresponding change in reachable positions given by
\[\mathcal{R}(P_t,E_t) \rightarrow \mathcal{R}(P_t,E_{t+1}).\]The resulting difference is not necessarily an immediate change in position. Instead, the environment may change which subsequent positions can be reached from the current position.
This gives environmental change a distinctive strategic character. An agent may choose to change its environment not merely to improve its current circumstances, but to alter the set of future positions that it can reach.
The resulting decision is therefore not simply whether an environmental change produces an immediate improvement. It is whether the change produces a sufficiently favorable alteration in future possibilities to justify the costs and risks involved.
This leads directly to the problem of environmental access.
The preceding analysis assumes that an environment $E$ is given. An agent occupies an environment, and the actions, costs, and consequences associated with that environment determine the positions available to it. But an environment is not necessarily accessible to an agent simply because it exists.
Let
\[\mathcal{E}=\{E_1,E_2,\ldots,E_m\}\]denote the set of environments that exist within the relevant game. The environments contained in $\mathcal{E}$ may differ substantially in the opportunities, resources, constraints, risks, and future positions they make available. The agent, however, may be able to access only a subset of these environments from its current position.
Let
\[\mathcal{A}_E(P_t)\subseteq\mathcal{E}\]denote the set of environments accessible to the agent from position $P_t$.
The distinction between the environment space and the accessible environment set is fundamental:
\[E\in\mathcal{E} \not\Rightarrow E\in\mathcal{A}_E(P_t).\]An environment may therefore exist and even provide access to highly desirable future positions while remaining inaccessible to the agent.
Environmental access depends on the conditions possessed by the agent. The resources, capabilities, knowledge, relationships, and other properties of the agent’s current position can determine which environments are accessible:
\[\mathcal{A}_E(P_t)=G(P_t),\]where the particular form of $G$ depends on the game. In one context, access to an environment may depend primarily on physical mobility. In another, it may depend on financial resources, knowledge, language, social connections, technological capabilities, or legal authorization. Often several such conditions must be satisfied simultaneously.
This means that access to an environment can itself be understood as a capability-dependent problem. An agent may possess the desire or preference to enter an environment without possessing the capabilities required to do so.
For example, an agent may have sufficient financial resources to benefit from an environment but lack transportation to reach it. An agent may have the physical ability to reach a location but lack the language required to operate effectively within it. An agent may possess both the resources and the means of transportation but lack the legal authorization required to enter or remain there.
These are not merely different examples of environmental constraints. They represent different components of the conditions required to make an environment accessible.
The problem therefore extends beyond choosing an environment. The agent may first have to change its own position in order to make that environment accessible.
An environmental transition
\[E_t\rightarrow E_{t+1}\]may consequently require an intermediate transition in position:
\[P_t \rightarrow P_{t+1} \rightarrow E_{t+1}.\]The first transition may involve acquiring the resources or capabilities necessary to make the second transition possible. In this sense, an agent can invest in environmental access by changing its current position in order to expand the set of environments available to it.
This produces an important distinction between using an environment and acquiring access to an environment. Acquiring access can matter even before the environment is actually entered, because access changes the set of future possibilities available to the agent.
If
\[\mathcal{A}_E(P_{t+1}) \supset \mathcal{A}_E(P_t),\]then the transition from $P_t$ to $P_{t+1}$ has expanded the set of environments accessible to the agent.
The expansion can matter because each newly accessible environment may provide access to a different set of future positions:
\[E\in\mathcal{A}_E(P_t) \rightarrow \mathcal{R}(P_t,E).\]Thus an improvement in position can have a second-order effect: it can increase not only the positions immediately available to the agent, but also the set of environments from which additional future positions can be reached.
This creates a recursive relationship between position and environmental access:
\[P_t \rightarrow \mathcal{A}_E(P_t) \rightarrow E \rightarrow \mathcal{R}(P_t,E) \rightarrow P_{t+k}.\]The environment problem therefore includes a problem of expanding environmental access. An agent may improve its position not only by pursuing a better position within its current environment, but by acquiring the capabilities necessary to access environments that provide more valuable future possibilities.
This introduces a strategic tradeoff. Resources used to expand environmental access cannot necessarily be used simultaneously to improve the agent’s current position. The agent must therefore evaluate whether an investment in access produces sufficient improvement in future possibilities to justify its immediate costs and risks.
The problem can be expressed generally as a choice between maintaining the current position and undertaking actions that expand the accessible environment set:
\[P_t \xrightarrow{\text{investment}} P_{t+1}\]such that
\[\mathcal{A}_E(P_{t+1}) \supseteq \mathcal{A}_E(P_t).\]The relevant question is therefore not simply which environment is best. It is which investments in environmental access produce the greatest improvement in the agent’s future possibilities relative to their costs, risks, and opportunity costs.
This is the beginning of an environmental strategy problem.
The problem of environmental access establishes that an agent does not necessarily have access to every environment that exists. The agent’s current position determines, at least in part, the environments available to it. This raises a further question: what happens when an agent deliberately changes the environment in which it operates?
An environmental change can be represented as
\[E_t\rightarrow E_{t+1}.\]Such a transition is distinct from an ordinary change in position within a fixed environment. An agent may move from
\[P_t\rightarrow P_{t+1}\]while remaining within the same environment $E_t$. Alternatively, the agent may undertake actions that change the environment itself:
\[E_t\rightarrow E_{t+1}.\]These two forms of change can have fundamentally different consequences. A change in position within an environment alters the agent’s state while leaving the surrounding conditions substantially unchanged. A change in environment can instead alter the conditions under which subsequent actions are taken.
The distinction can be represented as
\[(P_t,E_t) \rightarrow (P_{t+1},E_t)\]for a change occurring within an environment, versus
\[(P_t,E_t) \rightarrow (P_{t+1},E_{t+1})\]when the agent changes its environment.
The latter transition may alter the agent’s available actions, the costs of those actions, their consequences, and the environments that can subsequently be accessed. Consequently, an environmental change can produce effects that extend well beyond the immediate change in $P$.
In particular, an agent may undertake an action whose immediate effect on its position is neutral or negative because the action changes its future environmental possibilities. If
\[P_t\rightarrow P_{t+1}\]produces a temporary reduction in some component of position but results in
\[\mathcal{A}_E(P_{t+1}) \supset \mathcal{A}_E(P_t),\]then the agent has expanded the set of environments available to it.
This creates a distinction between the immediate return and the option value of environmental change. The immediate return concerns the change in the agent’s current position. The option value concerns the additional future possibilities created by the change.
The value of an environmental transition can therefore extend beyond the immediate difference
\[U(P_{t+1})-U(P_t).\]The transition may instead alter the set of future positions from which subsequent utility can be obtained:
\[\mathcal{R}(P_{t+1},E_{t+1}) \neq \mathcal{R}(P_t,E_t).\]This is particularly important when environmental changes are costly. An agent may need to spend money, time, knowledge, social capital, or other resources in order to change its environment. The relevant decision is therefore not whether environmental change is intrinsically beneficial, but whether the expected improvement in future possibilities justifies the resources and risks required to obtain it.
Environmental change can consequently be understood as an investment in the agent’s future decision space.
The agent may use resources available in its current position to acquire access to a different environment, with the expectation that the new environment will provide access to more valuable positions or opportunities. The transition can therefore take the form
\[P_t \xrightarrow{\text{investment}} P_{t+1} \rightarrow E_{t+1} \rightarrow \mathcal{R}(P_{t+1},E_{t+1}).\]The investment may be worthwhile even when the immediate position $P_{t+1}$ is not superior to $P_t$, provided that the expansion of future possibilities is sufficiently valuable.
This also means that environmental change need not always involve permanent relocation. An agent may temporarily enter another environment, maintain the ability to return to a previous environment, or acquire legal, financial, social, or technological capabilities that expand its access without immediately changing its physical location.
The relevant strategic object is therefore not simply the environment in which the agent is currently operating, but the relationship between the agent’s current position and the set of environments it can reach.
This leads naturally to the question of whether an agent should seek a single preferred environment or maintain access to multiple environments simultaneously. Once access to environments itself possesses value, maintaining alternative environments can become a source of optionality.
The problem therefore extends from environmental change to environmental optionality.
The problem of environmental access does not require an agent to choose a single environment and permanently commit to it. An agent may instead maintain access to multiple environments, each of which provides different opportunities, constraints, risks, and future possibilities.
Let
\[\mathcal{A}_E(P_t)\subseteq\mathcal{E}\]denote the set of environments accessible to an agent from position $P_t$. The agent’s environmental circumstances are therefore not characterized solely by the environment it currently occupies. They may also be characterized by the other environments to which it retains access.
Consider two agents occupying the same environment $E_1$:
\[E(P_1)=E(P_2)=E_1.\]Suppose, however, that their accessible environment sets differ:
\[\mathcal{A}_E(P_1)=\{E_1\}\]while
\[\mathcal{A}_E(P_2)=\{E_1,E_2,E_3\}.\]Although the agents currently occupy the same environment, they do not face the same environmental decision problem. The second agent retains the ability to enter environments that are inaccessible to the first.
This difference can affect the future positions available to the agents. For an environment $E_i$, let
\[\mathcal{R}(P_t,E_i)\]denote the positions reachable under that environment. The total set of future positions associated with the agent’s environmental options may therefore depend on the environments to which it retains access:
\[\mathcal{R}_E(P_t) = \bigcup_{E_i\in\mathcal{A}_E(P_t)} \mathcal{R}(P_t,E_i).\]Expanding environmental access can therefore expand the set of future positions available to the agent:
\[\mathcal{A}_E(P_{t+1}) \supset \mathcal{A}_E(P_t)\]may produce
\[\mathcal{R}_E(P_{t+1}) \supseteq \mathcal{R}_E(P_t).\]The additional environments need not be immediately used for this expansion to have value. An environment may be valuable because the agent retains the ability to enter it if circumstances change.
This creates a form of environmental optionality. The agent preserves future choices by maintaining the ability to operate in environments that it does not currently occupy.
The value of this optionality becomes particularly important under uncertainty. The future conditions of an environment may change, and the environment that is most favorable under current conditions may not remain so. Maintaining access to alternative environments can therefore reduce the consequences of an unfavorable change in the current environment.
The agent may consequently prefer a position with greater environmental optionality even when its immediate environment is identical to that of another agent.
This does not imply that every additional environment is valuable. Access may require resources to acquire, maintain, or exercise. An environment may also introduce obligations, risks, or opportunity costs. Environmental optionality therefore has a cost.
Let $\mathcal{S}\subseteq\mathcal{E}$ denote a set of environments to which the agent maintains access. The agent may face a tradeoff between the value of the future possibilities preserved by $\mathcal{S}$ and the resources required to maintain that access:
\[V(\mathcal{S}\mid P_t)-C(\mathcal{S}\mid P_t).\]The environmental problem therefore includes not only the question of which environment an agent should occupy, but also which environments it should retain the ability to access.
This distinction is important because environmental access can itself become a strategic asset. An agent may deliberately acquire or preserve access to an environment even when there is no immediate intention to use it, because doing so preserves the ability to respond to future changes in circumstances.
Environmental strategy therefore need not be a choice between environments at a single point in time. It may instead involve maintaining a set of environmental possibilities through time.
The problem that follows is consequently one of determining how an agent should allocate its limited resources among competing ways of acquiring, maintaining, and exercising environmental access.
The preceding sections establish that an agent’s environment is not merely a condition in which decisions occur. The environment affects the actions available to the agent, the costs and consequences of those actions, the environments that can subsequently be accessed, and therefore the future positions that can be reached.
The agent therefore faces a strategic problem: how should it allocate its limited resources among the actions that determine its environmental circumstances?
An agent may use its resources to improve its position within its current environment. Alternatively, it may use those resources to acquire access to another environment, change its environment, or preserve access to environments that it does not currently occupy. These alternatives compete for the same limited resources.
The choice can therefore be represented as
\[a_t\in\mathcal{A}(P_t,E_t),\]where the available actions include not only actions directed toward immediate improvements in position, but also actions that alter environmental access or environmental conditions.
An environmental strategy is consequently a policy governing how an agent responds to its environmental circumstances over time. We can represent such a strategy as
\[\pi_E(a\mid P_t,E_t),\]which specifies how the agent selects environmental actions as a function of its current position and environment.
The objective of the strategy is not necessarily to maximize the quality of the agent’s immediate environment. The relevant objective is to improve the agent’s ability to reach desirable future positions.
A general environmental strategy can therefore be expressed as an optimization problem:
\[\pi_E^* = \arg\max_{\pi_E} \mathbb{E}\left[U(P_T)\mid P_0,E_0,\pi_E\right].\]The expectation is important because the consequences of environmental decisions are rarely known with certainty. An environment that appears advantageous under current conditions may become less favorable. An attempt to acquire access may fail. The costs of maintaining access may change. Political, economic, technological, social, or physical conditions may also alter the relative value of different environments.
Environmental strategy is therefore a problem of decision-making under uncertainty.
This uncertainty also means that an agent need not always commit to a single environmental pathway. When several environmental strategies have uncertain outcomes, the optimal strategy may involve allocating resources among multiple alternatives rather than selecting one exclusively.
Let
\[\Pi_E=\{\pi_1,\pi_2,\ldots,\pi_k\}\]denote a set of available environmental strategies. An agent may adopt a mixed strategy
\[\sigma_E = (p_1,p_2,\ldots,p_k),\]where
\[\sum_{i=1}^{k}p_i=1\]and $p_i$ represents the probability assigned to strategy $\pi_i$.
The significance of a mixed strategy is that environmental uncertainty can make flexibility itself valuable. An agent that commits entirely to one environmental pathway may obtain a high payoff if conditions develop favorably, but may become exposed if those conditions change. Maintaining multiple pathways can reduce dependence on any single environmental outcome.
This connects environmental strategy directly to environmental optionality. Optionality is not merely a desirable property of a position; it can be deliberately produced through strategic action.
An agent may therefore pursue actions whose immediate returns are modest because those actions preserve or expand future environmental choices. The relevant calculation is not simply
\[\Delta U_{\text{immediate}},\]but the expected change in the future decision space produced by the investment.
An environmental strategy can consequently involve several distinct objectives:
\[\text{Environmental Strategy} \rightarrow \begin{cases} \text{improve the current environment},\\ \text{acquire access to new environments},\\ \text{maintain existing access},\\ \text{preserve alternative pathways},\\ \text{reduce exposure to environmental risk}. \end{cases}\]These objectives may conflict. Resources devoted to acquiring access to another environment cannot simultaneously be used for every possible improvement within the current environment. Maintaining an alternative may have a continuing cost even when it is never exercised. Changing environments may produce short-term losses in exchange for greater future possibilities.
The environmental strategy problem is therefore fundamentally an allocation problem.
The agent must determine how much of its available resources should be devoted to its current environmental circumstances and how much should be devoted to changing, expanding, or preserving its future environmental possibilities.
This distinction is especially important because environmental improvements can operate at different time scales. Some actions produce immediate benefits. Others create capabilities that may not produce value until a future environmental transition becomes necessary or advantageous.
An agent may therefore rationally sacrifice some immediate utility in order to acquire a capability that substantially expands its future environmental options:
\[P_t \rightarrow P_{t+1} \rightarrow \mathcal{A}_E(P_{t+1}) \supset \mathcal{A}_E(P_t).\]The resulting position may not initially appear superior when evaluated only by its immediate circumstances. Its value may instead arise from the additional environments and future positions that have become accessible.
The environmental strategy problem can therefore be stated more precisely:
Given an agent’s current position, environment, resources, and uncertainty about future conditions, which actions should the agent take to acquire, maintain, change, or expand environmental access in order to improve its expected future possibilities?
This formulation transforms the environment problem from a descriptive observation into an actionable decision problem. The agent does not merely ask whether its environment is favorable. It asks what can be done about the environmental circumstances in which it finds itself, what those actions will cost, what risks they introduce, and what future possibilities they create.
The answer will differ across games and across agents. There is no universally optimal environment independent of the agent occupying it, just as there is no universally optimal environmental strategy independent of the resources, capabilities, constraints, and objectives of the agent.
What is universal is the structure of the problem: an agent must operate within an environment, its environment affects its future possibilities, and the agent may have to invest resources to acquire, preserve, or improve access to environments that provide more valuable possibilities.
The specific strategies that follow from this structure are therefore applications of the general environmental strategy problem. Migration, relocation, transportation, legal authorization, international mobility, technological access, and other forms of environmental access are not separate fundamental problems. They are particular instances of the same underlying problem.
This is where the general environment problem gives way to specific environmental strategies.
The environment $E$ need not be treated as a single undifferentiated variable. For many problems, it can be represented as a collection of environmental conditions:
\[E= \left( E^{\mathrm{phys}}, E^{\mathrm{inst}}, E^{\mathrm{econ}}, E^{\mathrm{social}}, E^{\mathrm{tech}}, E^{\mathrm{info}}, E^{\mathrm{intl}} \right)\]Each component can affect the agent’s decision problem in a different way. The resulting action space can therefore be written as
\[\mathcal{A} = \mathcal{A} \left( P, E^{\mathrm{phys}}, E^{\mathrm{inst}}, E^{\mathrm{econ}}, E^{\mathrm{social}}, E^{\mathrm{tech}}, E^{\mathrm{info}}, E^{\mathrm{intl}} \right)\]A change in any environmental component can consequently alter the actions available to the agent.
For the physical environment,
\[E^{\mathrm{phys}}_t \rightarrow E^{\mathrm{phys}}_{t+1}\]may change the cost or feasibility of physical actions:
\[c(a,E^{\mathrm{phys}}_t) \neq c(a,E^{\mathrm{phys}}_{t+1})\]For the institutional environment,
\[E^{\mathrm{inst}}_t \rightarrow E^{\mathrm{inst}}_{t+1}\]may change the set of permissible actions:
\[\mathcal{A}(P_t,E_t) \neq \mathcal{A}(P_t,E_{t+1})\]For the economic environment,
\[E^{\mathrm{econ}}_t \rightarrow E^{\mathrm{econ}}_{t+1}\]may change the resources required to reach a position:
\[c(a,E^{\mathrm{econ}}_t) \neq c(a,E^{\mathrm{econ}}_{t+1})\]For the social environment,
\[E^{\mathrm{social}}_t \rightarrow E^{\mathrm{social}}_{t+1}\]may change the relationships and opportunities available to the agent, producing:
\[\mathcal{A}_E(P_t) \rightarrow \mathcal{A}_E(P_{t+1})\]For the technological environment,
\[E^{\mathrm{tech}}_t \rightarrow E^{\mathrm{tech}}_{t+1}\]may expand the agent’s available actions:
\[\mathcal{A}(P_t,E_{t+1}) \supset \mathcal{A}(P_t,E_t)\]and therefore expand the reachable set:
\[\mathcal{R}(P_t,E_{t+1}) \supseteq \mathcal{R}(P_t,E_t)\]The information environment can similarly affect the agent’s ability to identify and evaluate available actions. If $I_t$ represents the information available to the agent, then:
\[I_t \rightarrow I_{t+1}\]can change the agent’s effective decision problem even when its physical position and other environmental conditions remain unchanged.
Finally, the international environment illustrates the problem of environmental access particularly clearly. Changes in legal or political conditions can alter the set of environments accessible to an agent:
\[\mathcal{A}_E(P_t) \neq \mathcal{A}_E(P_{t+1})\]Across these cases, the particular mechanism differs, but the underlying structure remains:
\[E \rightarrow \mathcal{A}(P,E) \rightarrow \mathcal{R}(P,E)\]Environmental differences therefore need not produce their effects through the same mechanism. They may change what an agent can do, what an action costs, what consequences an action produces, or which environments the agent can subsequently access.
The general environmental problem is consequently applicable whenever variation in external conditions produces variation in the agent’s feasible actions or future reachable positions.
Thus the problem is not specific to geography, relocation, or migration. These are particular instances of a more general relationship between an agent and the environments available to it.
The remaining question is therefore not whether environment creates a strategic problem, but how that general problem should be addressed in specific circumstances.
The environment problem can now be stated in terms of the relationship between an agent’s position, its environment, and the environments accessible to it.
For an agent at position $P_t$ operating within environment $E_t$,
\[E_t \rightarrow \mathcal{A}(P_t,E_t) \rightarrow \mathcal{R}(P_t,E_t)\]determines part of the agent’s future decision space.
At the same time, the agent’s position determines, at least in part, which environments it can access:
\[P_t \rightarrow \mathcal{A}_E(P_t).\]The agent can therefore influence its future possibilities through two related mechanisms:
\[P_t \rightarrow \begin{cases} \text{actions within }E_t\\ \text{access to other environments} \end{cases}\]The second mechanism extends the decision problem beyond optimization within the current environment. An agent may invest resources to change the conditions under which it acts, acquire access to another environment, or preserve access to multiple environments.
The resulting strategic problem is therefore:
\[\pi_E^* = \arg\max_{\pi_E} \mathbb{E} \left[ U(P_T) \mid P_0,E_0,\pi_E \right].\]The optimal strategy depends on the agent’s position, available resources, accessible environments, costs, risks, and uncertainty about future conditions. Consequently, there is no universally optimal environmental strategy independent of the agent and the game in which it operates.
The general problem can instead be understood as the management of environmental possibilities:
\[\boxed{ \text{Acquire} \rightarrow \text{Maintain} \rightarrow \text{Expand} \rightarrow \text{Exercise} }\]the agent’s access to environments that improve its future possibilities.
Different mechanisms of environmental access produce different strategic problems. Transportation, financial resources, social connections, technology, legal authorization, and other capabilities can each determine whether an environment is accessible and how effectively an agent can operate within it.
These specific mechanisms can therefore be analyzed independently while retaining the same underlying framework.
The environment problem is consequently the general problem. Specific environmental strategies are its applications.
The next step is to examine those applications directly.