Role Allocation in a Group of Control Objects: General Purpose Approach

Role Allocation in a Group of Control Objects: General Purpose Approach

Viacheslav Abrosimov (Smart Solutions, Russia)
Copyright: © 2017 |Pages: 19
DOI: 10.4018/978-1-5225-2322-2.ch010

Abstract

The efficiency of control objects that fulfill economic and military tasks in groups depends on the correct role allocation among them. This paper develops a role allocation algorithm for control objects jointly fulfilling a collective task. Control objects are represented as intelligent agents with some capabilities and needs. The problem is solved by ranking the agents based on their closeness to given roles in terms of their functionality and characteristics. The efficiency of the suggested approach is illustrated by allocating the role of reconnaissance aircraft in the group attack problem with a protected target.
Chapter Preview
Top

Introduction

While considering practical problems related to the operation of control objects in a complex conflict environment, researchers pay much attention to the design of appropriate solutions not for single control objects but for their groups. It is well-known that such groups solve tasks with higher reliability and speed, mainly owing to the mutual support of the control objects and their capability for self-organization (e.g., see Rzevski & Skobelev, 2014). The efficiency of a group depends on role allocation within it, whether the group is guided by a leader or has self-organization. Role allocation concerns role assignment, the delegation of authorities to different roles, etc.

As a matter of fact, there exists no clear definition of role in technical systems. Each object occupies a definite position in a group hierarchical structure. This position predetermines the object’s status in the group. Role actually describes the functionality that must be implemented by an object while fulfilling a collective task. Role and status are closely associated with each other. Generally, in hierarchical structures a status defines a corresponding role. The self-organizing structures without fixed hierarchies have the opposite picture, i.e., the status of an object is defined after role allocation.

The primary objective of group formation is collective goal setting. As a rule, a collective goal can be decomposed into special tasks. Roles are directly defined by the range of tasks allocated (in hierarchical structures) or undertaken (in self-organizing structures) by a control object. The whole meaning of role consists in the following. Being selected, allocated or accepted for execution by a control object (and further reported to all objects of a group), a role allows this object to predict the behavior of other objects and also the behavior of the whole group; based on roles, other control objects plan their behavior and actions.

Control objects have definite functionality and resources, thereby pretending to the receipt and further execution of roles. There are two fundamentally different ways to occupy a role in a group, namely, (a) assignment (e.g., by Leader) and (b) selection on a competitive basis. The second case is of special interest. In mathematical terms, the relations of control objects during competitive role allocation within a group should be considered using the multiagent approach (Weiss, 2013). Really, the representation of control objects as intelligent agents seems fruitful (Abrosimov, 2015). Intelligent agents may possess properties that are inherent in intelligent control systems, viz., rationality, existing goals and knowledge, and autonomy. In contrast to social agents (whose role functions are mostly connected with the notion of agents interaction for social behavior analysis [Trzebiatowski & Münch,]), here the matter concerns the agents that describe technical systems. For such agents, the well-known belief-desire-intention (BDI) concept is formulated via the notions of capability and need (Rzevski & Skobelev, 2014). Therefore, the capabilities of modern control objects can be expressed as the capabilities of agents, the design features of control systems as the properties or agents and the functional specifics (including constraints) as the needs of agents. This scheme well fits the competition of the agents during role occupation, particularly in the context of self-organization.

In the sequel, we consider the control objects that are representable and formalizable as intelligent agents.

Complete Chapter List

Search this Book:
Reset