Complex decisions can be understood as an unpredictable process, uncommon and composed of actions. To handle this complexity, it is necessary to think and apply strategies to mitigate the risks and the unpredictable actions concerned to a decision making. Design the complex decision impact can bring alternatives to solve problems with less damage to the environment and humans involved. In this paper, we propose an impact projection's conceptualization, modelled into an ontology called OntoImpact*. This ontology proposal has a differential in representing the state of art about complex decisions impact, through the decision-making process, impact characterizers, agents involved and contextual information. This ontology is based on the unified foundational ontology (UFO) and is organized into five aspects: decision agents and domain characteristics; collaboration aspects; complex decision process and activities; structural elements of impact projection; and impact details. In this paper we present the contribution of UFO in the meaning of OntoImpact*.
Decision-making is a common process in humans’ life. The more diverse the environment in which
the decision is made, the complex are the tasks. Every day, decision makers face decision tasks that are
chaotic, complex, and interrelated [
Impact projection is a natural initiative in decision making field. However, it is a tacit or individual
task. Some research [
This research extends our previous work [
The main contribution of this paper is the OntoImpact*. It is an ontology, with new theoretical discussions on its semantic expressiveness. This ontology comprises concepts from several perspectives related to decision impact, and it is organized in five modules: decision agents and domain characteristics; collaboration; process and activities; the structural elements; and impact details.
In practice, OntoImpact* is intended to form the basis to analyze a set of complex decisions from the perspective of their impacts, in a deep way. This ontology plays the role to externalize the concepts and its relations, able to influence decision analysts on their analysis, especially the decisions classified as naturalistic decision making (NDM). In this paper, we present the ontology`s development process, highlighting the concepts, its origins, relationships, and the semantic improved by UFO.
The paper is organized as follows: Section 2 presents fundamental concepts regarding the impact projection of complex decisions and related work; Section 3 presents OntoImpact*; Section 4 analyses OntoImpact*; and Section 5 concludes the paper and highlights future directions.
An impact is a measure of the tangible and intangible consequences. Project impacts involves
making predictions of potential future outcomes [28]. It is possible to find works in the literature that
discuss the impact of a decision in the planning phase of the decision process [
The decision support and impact projection approaches found in the literature include smart choice
[
Decision impacts are considered in different ways in seven of eight approaches verified. The decision-making approaches in Table 1 support impact discussion. Despite this, little work has been done on systematizing the impacts projection of complex decisions still in the planning phase of the Technology ✔ ✔ ✔ decision process, considering the cognitive aspects of the decision-makers involved and facing lack or ambiguous information about the decision.
In view of these considerations, we propose a formal conceptualization for the impact projection of
complex decisions, based on the characterization elements mentioned above and on the existing
understanding of impact projection. This conceptualization is represented in the form of OntoImpact*,
an ontology based on UFO [
An ontology represents an abstraction of reality and explicit a shared conceptualization [
OntoImpact* involves the search for existing ontologies related to domain and scope that were
established in the previous steps of development of this ontology. Two proposals of interest for reuse
were found in the development of OntoImpact*: the knowledge-intensive process ontology (KIPO) [
The cognitive aspects and competencies of decision makers related to impact projection can be seen in Figure 1. The impact projection for a complex decision involves several agents that may be specialists or decision makers, with specific backgrounds and rationales. 3.2.
Collaboration is an important aspect to improve exchange and development of knowledge. In impact projection of complex decision, it is a determining factor in the analysis, discovery, and implementation of new tasks. In the domain modelled by OntoImpact*, socialization is a central interaction that permits discussion and analysis of the main concepts, such as decision alternatives, the decision scenario, and impacts. Figure 2 presents the Collab, providing a shared understanding of how collaboration can improve impact projection of complex decisions.
The decision maker group is composed of decision makers who socialize to discuss the actions inherent in the impact projection activity. It is the role of the decision-maker to analyze the impacts of socialization and to use the available information resources to support the impact projection activities of the decision. 3.3.
This module of OntoImpact* concerns the impact projection process and the elements that constitute its activities. This module is concerned to understand the dynamic of complex decision and for that discussed the influences of the different types of goals and the activities involved.
IP module (Figure 4) presents the structural elements of impact projection and the projection mechanism to project an impact beyond collaborative actions. OntoImpact* present the relations concerned to the concrete concepts like alternative and activities of impact projection, with cognitive aspects.
The ID module describes the concept of an impact and its interactions in the context of complex decisions. This module answered questions related to specificities of impact. For that, this module discussed the relations between events, type of events, risk, and vulnerability for instance.
OntoImpact* defines a domain of high complexity that involves the projected impacts of a complex decision dynamics. To achieve results with less ambiguity in terms of the domain, and a better understanding of the impacts of a complex decision and its associated types, this ontology was built based on the methodology and competence questions discussed above, in addition to UFO fundamentals represented into the stereotypes presented in the Figures 1-5.
A UFO is divided into three layers: UFO-A, UFO-B and UFO-C. UFO-A defines the core of UFO,
and involves what we call endurants. UFO-B involves perdurants. The difference between these lies in
their behavior over time: endurants persist throughout time, such as people and buildings, while
perdurants are composed of temporal parts over a period of time. A perdurant can be understood as an
event, such as a process, a meeting or a football match. UFO-C involves social entities, such as agents,
and their behavioral. It is based on UFO-A and UFO-B. During the development of OntoImpact*, 23
constructs were used from UFO layers. Four constructs from UFO-A were used, based on stereotype,
and 19 constructs from UFO-B and UFO-C. The discussion below reflects the formalization of UFO
based on several reference elements [
UFO-A involves two categories: universals and particulars. Each particular is understood as an instance
of some universal and is also an entity that exists in reality with a unique identity. A universal is defined
as a set of characteristics that can be perceived in different individuals. Individuals can assume two
types: concrete particular and abstract particular. Concrete individual types are endurant, and are
categorized into three types: substantial, mode (or moment), and situation. A situation endurant is an
element that represents complex entities made up of several objects or other situations. In this way, the
current situation represents a portion of reality that can be understood as a whole [
The universals act as groups of individuals. There are two types of universals in the UFO-A category:
substantial and moment universals. The substantial universals are groups of substantial individuals, and
moment universals are groups of moment individuals. UFO-A defines [
Sortal universals are classified into rigid, non-rigid or anti-rigid. Rigid sortals bring the rigid semantic charge, that is, an individual who possesses it must continue to have it as long as it exists. For an anti-rigid sortal, the properties of individuals do not necessarily apply to all their instances. The nonrigid sortal is a universal that does not necessarily apply to at least one of its instances. A rigid sortal is divided into four types: kind, subkind, collective and quantity. Kind is understood as a sorcery substance, and its instances are complex natural and artificial functions, such as person, company, desk or computer. The phase and role types are anti-rigid, where a role corresponds to an instantiation of an event or relation participation.
Figure 1 illustrates the concept of a kind in the form of an external agent. A group of decision makers interacts with the outside agent, who is a kind. The external agent presents the same principle of identity and may be another decision maker, expert, or certain objects that are able to provide new inputs to be analyzed. OntoImpact* externalizes role-playing in the interaction between decision makers for the action’s execution focused on the impact projection. Figure 2 shows the sender and receiver roles that can be played by a decision maker.
The construct category is a type of rigid mixin that is a universal able to join properties common to different substance sortals, such as the resource concept shown in Figure 1. Here, a resource that is a category can be date or time type. A roleMixin is an externally dependent anti-rigid nonsortal that aggregates properties common to different roles. Figure 4 shows an example of a reused roleMixin from the OntoEmerge ontology, which is the impacted environment concept.
UFO-A also discusses the concepts of first order universal (FOU) and high order universal (HOU). The former encompasses universals such as a person or animal, whose instances are individuals, while the latter represents universals that have FOU instances. An example of HOU would be a type of tool, for which the instances would be hammer and a sledgehammer. Figure 5 gives examples of HOUs, such as a hazardous event, improvement event, neutral event, improvement situation, and impact. Figure 5 presents concepts defined as HOUs: the type of alternative, and the unknown, known, discarded and chosen alternatives. 4.2.
The UFO-B is dedicated to investigating events (perdurants) and enduring individuals (endurants).
Endurants are characterized by being always the same individual, whereas events are composed of
temporal parts such as a conversation or a game. In [
As events, they transform one situation (presented in UFO-A) into another, altering the state of the represented domain. Events are entities that are existentially dependent on their participants, since they would not exist if there was no participation by substantials. In UFO-B, events can be complex or atomic. Complex events are composed of at least two other events, that are either atomic or complex. In them, their parts add identity to the complex event. Atomic events are indivisible, according to the domain semantics. Examples of the application of an atomic event construct can be observed in Figure 5 through the concepts of hazardous, improvement and neutral events. Each of these contributes to the occurrence of a complex event that may be a damage, improvement, or neutral event. The occurrence of events may enable the manifestation of dispositions in certain situations. In general, objects have properties, some of which are defined as dispositions. A situation allows an event to occur when it activates the mood that is manifested by this event. Figure 5 shows the occurrence of a situation and disposition. Here, a vulnerability is a disposition manifested by the occurrence of an impact event. From this figure, it can be observed that the identified risk, which is an arrangement, plays the role of generating impact that is an event. 4.3.
UFO-C focuses on social concepts involving actions, agents, an intention, a plan, and a commitment. UFO-C has a structure that is based on UFO-A and UFO-B.
Agents in UFO-C can assume one of three types: physical, social or society. In the same way, objects can be physical or social objects. Examples of social objects are organizational rules and norms. A normative description is a type of social object that is able to define rules that are known by at least one social agent. In Figure 3, the impact report is a normative misrepresentation that must be recognized by the organization, in this case behaving as a social object.
Agents are substantial types that may have modes called intentional moments. Each intentional moment has a single proposition as its propositional content. An intention is a subtype of a mental moment, which is a subtype of an intentional moment. From Figure 4, we can observe the occurrence of a proposition assumed by a message; this communicative act has propositional content. When a communicative interaction occurs, it is composed of a communicative act and a perception of this communicative act. Figure 5 shows that agents perform a collaborative session through action contributions, as exemplified through the concept of perception.
Concepts such as belief, desire, and intention (BDI), as discussed in [31], are considered in UFO to be types of mental moment and to inherit the characteristics of the intentional moment that is inherent to each agent. A mental moment is an intrinsic moment that is existentially dependent on a particular agent and is an inseparable part of the agent's mental state. From Figure 1, it can be observed that the concepts of culture, experience, preconceptions, creativity, and feeling are mental moments assumed by agents in the execution of their impact projection actions. In addition to these, agents also assume BDI mental moments in the execution of actions.
Actions and events are distinct elements in UFO-C. An action is caused by the intention of an agent, whereas an event requires the participation of a substantial. It is worth noting that an event is not initiated by participation. In UFO, an agent performs an action based on its commitment to fulfill an intention, with the activity goal as its propositional content. This assumption is illustrated in the AC module of OntoImpact* (Figure 1), in which an agent playing the decision-making role performs impact projection activities based on an intent to achieve the objective of the activity. In an impact projection for a complex decision, decision makers execute a projection activity, which is an action involved in the impact projection process.
An action may be complex or atomic. A complex action is composed of two or more participations, unlike an atomic action. One type of complex action is interaction, which is composed of contributions from different agents, for example in the dialogue between decision makers. An example of this assertion is observed in Figure 2. For socialization to take place, the participation of a decision maker group is required, composed of decision makers playing the roles of sender and receiver. Socialization involves the semantics of the interaction concept.
OntoImpact* covers the relevant perspectives for a complete conceptualization of impact projection for complex decisions, based on UFO. The main contribution is related to support decision makers into their works tasks, specifically the one that involves decision impacts.
Future work includes the evaluation of OntoImpact*, with special interest in analyzing its
completeness and usability to cover the main aspects of impact projection of complex decision. To
improve the collaborative session, we plan to develop a technology environment, based on
OntoImpact*, to support the impact projection of a complex decision. It is also a future work, improve
the OntoImpact* conceptualization through the Core Ontology on Decision Making [
This research was partially supported by the National Council for Scientific and Technological Development (CNPQ) under grant # 308149/2015-7, and by Rio de Janeiro Research Support Foundation (FAPERJ) under grant # E-26/202.876/2018 and # E-26/211.367/2019 (248406). [27] Orasanu, J., Connolly, T.: The Reinvention of Decision Making. In: Klein, G.A., Orasanu, J., Calderwood, R., Zsambok, E. (eds.) Decision Making in Action: Models and Methods, pp. 3–20.
Ablex Publishing Corporation, Norwood, New Jersey (1993) [28] Pant, P. N. and Starbuck, W. H.: Innocents in the forest: Forecasting and research methods. Journal of Management, 16: 433–460 (1990) [29] Pereira, A.C.T.D., Santoro, F.M.: Cognitive Decision Making Process as Context Information. In: Proceedings of the 15th IFIP WG8.3 International Conference on Decision Support Systems, Lisboa (2010) [30] Proctor, R.W., Zandt, T.V.: Human Factors in Simple and Complex Systems. Third edition. Taylor & Francis Group (2018). ISBN 9781315156811 [31] Rao, A. S., Georgeff, M. P.: Modeling rational agents within a BDI architecture. Second International Conference on Principles of Knowledge Representation and Reasoning (KR’91), pp. 473–484. Morgan Kaufmann Publishers, Cambridge, MA, USA (1991) [32] Rasmussen, J.: Information Processing and Human-Machine Interaction: An Approach to
Cognitive Engineering. North-Holland, New York (1986) [33] Shattuck, L.G., Miller, N.L.: Extending naturalistic decision making to complex organizations: A dynamic model of situated cognition. Organization Studies 27, 989–1009 (2006) [34] Shortland, N. D., Alison, L. J., Moran, J. M.: Conflict: How Soldiers Make Impossible Decisions.
Oxford University Press, pp. 240, 2019. [35] Stanton, N.A., Bessell, K.: Team decision making: Ethe structure of interdependences when returning to periscope depth. In: International Conference on Naturalistic Decision Making. Marseille, France (2013)