2.2.2.1. Telecommunications networks

The world of telecommunication networks is extensively modeled on the basis of this concept. A telecommunication network is a mesh of nodes fulfilling a variety of tasks. Each node is an agent. It can be defined as a computational entity:

*acting on behalf of other entities in an autonomous fashion (proxy agent);

*performing its actions with some level of proactivity and/or reactiveness;

*exhibiting some level of the key attributes of learning, cooperation and mobility.

Several agent technologies are operated mainly in the telecommunications realm. They fall into two main categories, i.e. distributed agent technology and mobile agent technology.

Distributed agent technology refers to a multi-agent system described as a network of actants with the following advantages:

• – solving problems that may be too large for a centralized agent;
• – providing enhanced speed and reliability;
• – tolerating uncertain data and knowledge.

They include the following salient features:

• – communicating between themselves;
• – coordinating their activities;
• – negotiating their conflicts.

“Actants” are non-human entities such as configurations of equipment, mediators and software programs and are distinguished from actors that are human beings. But actors and “actants” are entangled in ways that provoke complexity dynamics in many circumstances.

Mobile agent technology functions by encapsulating the interaction capabilities of agents into their descriptive attributes. A mobile agent is a software entity existing in a distributed software environment. The primary task of this environment is to provide the means which allow mobile agents to execute. A mobile agent is a program that chooses to migrate from machine to machine in a heterogeneous network.

The description of a mobile agent must contain all of the following models:

An agent model (autonomy, learning, cooperation).

A life-cycle model: this model defines the dynamics of operations in terms of different execution states and events, triggering the movement from one state to another (start state, running state and death state).

A computational model: this model, being closely related to the life-cycle model, describes how the execution of specified instructions occurs when the agent is in a running state (computational capabilities). Implementers of an agent gain access to other models of this agent through the computational model, the structure of which affects all other models.

A security model: mobile agent security can be split into two broad areas, i.e. protection of hosts from malicious agents and protection of agents from hosts (leakage, tampering, resource stealing and vandalism).

A communication model: communication is used when accessing services outside of the mobile agent during cooperation and coordination. A protocol is an implementation of a communication model.

A navigation model: this model concerns itself with all aspects of agent mobility from the discovery and resolution of destination hosts to the manner in which a mobile agent is transported (transportation schemes).

2.2.2.2. Manufacturing scheduling

Scheduling shop floor activities is a key issue in the manufacturing industry with respect to making the best economical use of manufacturing equipment and bringing costs under control, as well as delivering committed customer orders at due dates.

Consider the product structure from the manufacturing point of view as portrayed in Figure 2.1.

P_i are parts machined in dedicated shops and A_i are assembled products. In terms of scheduling, the relevant combined attributes of the products and equipment units involved in machining and assembling, whatever their layout job shop or batch or continuous line production, are lead times, so that a Gantt chart can be derived by backward scheduling from the due date when the “root” product A₄ or a batch thereof has to be delivered to its client. The situation is pictured in Figure 2.2.

Within the framework of decentralized decision-making for making collaborators more motivated for solving problems, they have to deal with and deploy what is called “job enrichment”; the choice of three agents, J₁ in charge of the delivery of A₂, a second one J₂ in charge of the delivery of A₃, and a third one J₃ in charge of the delivery of A₄ to the final client, appears to be the most pragmatic and efficient solution. This third agent is in some way the front office of all the hidden backward activities and responsible for fulfilling the commitments taken with respect to clients.

The three agents J_i have to collaborate for establishing a schedule over a time horizon commensurate with the lead times. When manufacturing problems of any sort, which can impact the fulfilment of the schedule, the agents J_i have to collaborate to devise a coherent, sensible solution (outsourcing, hiring extra workforce, etc.), often without letting the top management know the details of the problems but only the adequate courses of action taken.

2.4.2.1. Definition of belief

The verb “believe” is defined in the Oxford dictionary. We are aware that the word “belief ” has been interpreted in different ways in the realm of philosophy and that its translation in other Indo-European languages (German, French, Italian, Spanish, among others) appears difficult (Cassin 2004). It is outside the scope of this book to discuss this issue. We will interpret it within the framework of what is called “the philosophy of mind” in the English language context. According to Hume’s book (1978 I sec 7), a matter of fact is “easily cleared and ascertained” and is closely correlated with reality: “if this be absurd in fact and reality, it must be absurd in idea”. These matters of fact are objects of belief: “it is certain that we must have an idea of every matter of fact which we believe… When we are convinced of any matter of fact, we do nothing but conceive it” (Hume 1978, I, III sec 8). In his book “Enquiries Concerning Human Understanding and Concerning the Principles of Morals”, Hume (1975) confirms that matters of fact and relations of ideas should be clearly distinguished: all the objects of human reason or enquiry may naturally be divided into two kinds, namely, relations of ideas or matters of fact.

Some people distinguish dispositional beliefs and occurring beliefs to try to mirror the storage structures of our memory organization. A dispositional belief is supposed to be held in the mind but not currently considered. An occurring belief is a belief being currently considered by the mind.

2.4.2.2. Attitudes and beliefs

An attitude is a state of mind favorable to behave in a positive or negative way towards an object or a subject. The information–integration tenet is one of the best credible models of the nature of attitudes and attitude change as stated by Anderson (1971), Fishbein (1975) and Wyer (1974). According to this approach, all pieces of information have the potential to affect one’s attitude. Two parameters have to be considered to understand the degree of influence information has on attitudes, i.e. the how and how much parameters. The how parameter is intended to evaluate the extent to which a piece of information received supports one’s belief. The how much parameter tries to measure the weight assigned to different pieces of information for impacting one’s attitude through a change in one’s belief.

Attitudes are dependent on a complex factor involving beliefs and evaluation. It is important to distinguish between two types of belief, i.e. belief in an object and belief about an object. When one believes in an object, one predicts a high probability of the object attributes existing. Belief about is the predicted probability that particular relationships exist between one object and others. Beliefs are embodied by the hundreds of thousands of statements we make about self and the world.

Attitudes change when beliefs are altered when acquiring new knowledge. The quantitative assessment of an attitude towards an object or a subject is measured in terms of the weighted sum of each belief about that object or subject times its circumstanced valuation. M. Rokeach has developed an extensive explanation of human behavior based on beliefs, attitudes and values (Rokeach 1969, 1973). According to him, each person has a highly organized system of beliefs, attitudes and values, which guides behavior. From M. Rokeach’s point of view, values are specific types of beliefs that act as life guidance. He concludes that people are guided by a need for consistency between their beliefs, attitudes and values. When a piece of information brings about changes in attitude towards an object or a situation, inconsistency develops and creates mistrust.

Another facet of belief and trust is linked to certainty and probability. Probability is commonly contrasted with certainty. Some of our beliefs are entertained with certainty, while there are others of which we are not sure. Furthermore, our beliefs are time-dependent along with our acquaintanceships.

2.4.2.3. Beliefs and biases

Biases are nonconscious drivers, cognitive quirks that influence how people perceive the world. They appear to be universal in most of humanity, perhaps hardwired in the brain as part of our genetic heritage. They exert their influence outside conscious awareness. We do not take action without our biases kicking in. They can be helpful by enabling people to make quick, efficient judgments and decisions with minimal cognitive effort. But they can also blind a person to new information or inhibit someone from considering valuable data when taking an important decision.

Biases often refer to beliefs that appear as the grounds on which decisions and courses of action are taken. Below is a list of biases commonly found in social life:

• – in-group bias: perceiving people who are similar to you more positively (ethnicity, religion, etc.);
• – out-group bias: perceiving people who are different from you more negatively;
• – belief bias: deciding whether an argument is strong or weak on the basis of whether you agree with its implications;
• – confirmation bias: seeking and finding evidence that confirms your beliefs and ignoring evidence that does not;
• – availability bias: making a decision based on the data that comes to mind more quickly rather than on more objective evidence;

• – anchoring bias: relying heavily on the first perception or piece of information offered (the anchor) when considering a decision;
• – halo effect: letting someone’s positive qualities in a specific area influence the free will of one individual or a group of individuals (constraints, lobbying, etc.);
• – base rate fallacy: when judging how probable an event is, ignoring the base rate (overall rate of occurrence);
• – planning fallacy: underestimating how long a task will be taken to complete, how much it will cost, i.e. the risks incurred, while overestimating the benefits;
• – representativeness bias: believing that something that is more representative is necessarily prevalent;
• – hot hand fallacy: believing that someone who was successful in the past has a greater chance of achieving further success.

2.4.2.4. Degrees of belief

2.4.2.5. Belief, trust and truth

Truth is an attribute of beliefs (opinions, doctrines, statements, etc.). It refers to the quality of those propositions that accord with reality, or with what is perceived as reality. The contrast is with falsity, faithlessness and fakery.

Many explanations have been devised to elaborate on the correspondence between what is true and what makes it true. The correspondence theory asserts that a belief is true provided that a fact corresponding to it exists. What does it mean for a belief to correspond to a fact? How to verify that a fact exists in the context of virtual reality? A third party, trust, seems adequate to intervene within this framework to assess the credibility of information sources. The state of believing involves some degree of confidence towards a propositional object of belief.

Other theories have been proposed to explain how a belief is accepted as true. The coherence theory developed by Bradley and Blanchard asserts that a belief is verified when it is part of an entire system of beliefs that is consistent and harmonious. A statement S is considered logically true if and only if S is a substitution-instance of a valid principle of logic. The pragmatic theory produced by two American philosophers C.S. Pierce and W. James asserts that a belief is true if it works, if accepting it brings success.

In a book about the impact of blockchain technology on business operations (buying and selling goods and services and their associated money transactions), Don and Alex Tapscott (2016) estimate that a trust protocol has to be established according to four principles of integrity:

• – Honesty has become not only an ethical issue but also an economic one. Trusting relations between all the stakeholders of business and public organizations have to be established and made sustainable.
• – Consideration means that all parties involved respect the interests, desires or feelings of their partners.
• – Accountability means clear commitments and abiding to them.
• – Transparency means that information pertinent to employees, customers and shareholders must be made available to avoid the instillation of distrust.

This protocol shows how social actors have become aware of the importance of societal relationships in a faceless virtual world.

Blockchain is a distributed ledger technology. Blockchain transactions are secured by powerful cryptography that is considered unbreakable using today’s computers.

We resort to K. Lewin’s field theory to analyze how emotions, feelings, beliefs, truth and trust are dynamically articulated when agents interact and perform their activities (Lewin 1951). The fundamental construct introduced by K. Lewin is that of “field”. K. Lewin gained a scientific background in Germany before immigrating to America. That explains why he was led to introduce the concept of a “field” to characterize the spatial–temporal properties of a human ecosystem. This concept is widely used in physics to describe the physical properties of phenomena in a limited space.

All behavior in terms of actions, thinking, wishing, striving, valuing, achieving, etc. is conceived of as a change in some state of a “field” in a given time unit. Expressed in the realm of individual psychology, the field is the life space of the individual (Lebensraum in German culture). The life space is equipped with beliefs and facts that interact to produce mental states resulting in attitudes at any given time. K. Lewin’s assertion that the only determinants of attitudes at a given time are the properties of the field at the same time has caused much controversy. But it sounds reasonable to accept that all the past is incorporated into the present state of the field under consideration. To put it in a different wording, only the contemporaneous system can have effects at any time. As a matter of fact, the present field has a certain time depth. It includes the “psychological” past, “psychological” present and “psychological” futures which constitute the time dimension of the life space existing at a given time.

This idea of time dimension is also found in the concept developed by Markov to approach the description of stochastic processes by chain of events. State changes in a system that occur in time follow some probability law. The transition from a certain state at time t to another state depends only on the properties of the state at time t: all the past features of previous states are considered already included in the attributes of the present state.

All attitudes depend on the cognitive structure of the life space that includes, for each agent of a cluster, the other stakeholders of the cluster. When exposed to the behavior suggestions of other cluster agents or their critical judgment of his/her own behavior, every agent develops either a conditioned reflex based on his innate and/or acquired knowledge embedded in his/her brain’s neural connections or branches out into emotional expressions according to the way the received information is appraised as a reward or a threat. This last case occurs if (s)he feels (s)he cannot secure the right pieces of knowledge to produce an appropriate reaction. T.D. Wilson, D.T. Gilbert and D.B. Centerbar (2003) wrote “helplessness theory has demonstrated that if people feel that they cannot control or predict their environments, they are at risk for severe motivational and cognitive deficits, such as depression”.

If one organization agent trusts the other organization agents, his/her motivation is strengthened to embark on a learning process to better his/her acquired knowledge. Learning engages imagination, demands concentration, attention, efforts and trust in other agents’ good will. Conscious awareness is fully involved.

2.4.2.6. Beliefs and logic

Logic is the study of consistent sets of beliefs. A set of beliefs is consistent if the beliefs are not only compatible with each other but also do not contradict each other. Beliefs are expressed by sentences. When written, these sentences stating beliefs are called declarative.

Many sentences do not naturally state beliefs. One sentence may have different meanings or interpretations depending on the context. Beliefs are, in some way or another, the outcome of “rational” reasoning. By rational, it is meant that rules of logic are called on for justifying the conclusions reached. But which rules?

Classical logic can be understood as a set of prescriptive rules defining the way reasoning has to be conducted to yield coherent conclusions. Within this framework, truth is unique. It is implicitly assumed that a universe exists where propositions are either true or false. The “principle of the excluded third” is called on. It is well suited for data processing by computer systems. Data coded by binary digits are memorized and processed by electronic devices able to be maintained only in two states (0 or 1). Classical logic does not mirror the way we reason in our daily life. It is acknowledged that our brain does not operate as a Turing machine (Wilson 2003). If our brain is viewed as a black box converting input into output, the transformation process can be represented by algorithms. But the intimate physiological mechanism cannot be ascribed to algorithmic procedures in the way a computer system crunches numbers.

Other systems of logic, descriptive by nature, have been worked out to try to take into account the ways and means we use to make decisions in our daily activities. Modal logic is a system we practice, generally implicitly. Modality is the manner in which a proposition or statement describes or applies to its subject matter. Derivatively, modality refers to characteristics of entities or states of affairs described by modal propositions.

Modal logic (Blackburn 2001) is a branch of logic which studies and attempts to systematize those logical relations between propositions which hold by virtue of containing modal terms such as “necessarily”, “possibly” and “contingently”; must, may and can. These terms cover three modalities: necessity, actuality and possibility. In short, modal logic is the study of necessity (it is necessary that…) and possibility (it is possible that….). This is done with the help of the two operators □ and ◊ meaning “necessarily” and “possibly”, respectively, and instrumental in dealing with different conceptions of necessity and possibility:

• – logical necessity, i.e. true by virtue of logic alone (if P then Q);
• – contextual necessity, i.e. true by virtue of the nature and structure of reality (business context, social context, etc.);
• – physical necessity, i.e. true by virtue of the laws of nature (water boils at 100°C under standard pressure).

Modal logic is not the name of a single logical system; there are a number of different logical systems making use of the operators □ and ◊, each with its own set of rules.

Modal operators □ and ◊ are introduced to express the modes with which propositions are true or false. They allow logical opposites to be clearly elicited. The operators □ and ◊ are regarded as quantifiers over entities called possible worlds. □ A is then interpreted as saying that A is true in all possible worlds, while ◊ A is interpreted as saying that A is true in at least one possible world.

The two operators are, in fact, connected. To say that something must be the case is to say that it is not possible for it not to be the case. That is, □ A means the same as ¬◊¬A. Similarly, to say that it is possible for something to be the case is to say that it is not necessarily the case that it is false. That is, ◊A means the same as ¬□¬A. For good measure, we can express the fact that it is impossible for A to be true, as ¬◊A (it is not possible that A) or as □¬A (A is necessarily false). The truth value of ◊A cannot be inferred from the knowledge of the truth value of A. Modal operators are situation-dependent. Following the 17th Century philosopher and logician Leibniz, logicians often call the possible options facing a decision-maker possible worlds or universes. A fresh approach to the semantics theory of possible worlds was introduced in the 1950s by Kripke (1963a and 1963b).

To say that ◊A is true, it is required to say that A is in fact true in at least one of the possible universes associated with a decision-maker’s situation. To say that □A is true implies that A is true in all the possible universes associated with a decision-maker’s situation. The modal status of a proposition is understood in terms of the worlds in which it is true and worlds in which it is false. Contingent propositions are those that are true in some possible worlds and false in others. Impossible propositions are true in no possible world.

Two logical operators, i.e. negation and the conditional operator → (if …then …), which are central in decision-making, require special attention to be applied within the framework of possible worlds. Let us assume that a decision-maker is in a situation M and that M is a set of exclusive possible worlds. Each element of the set is a world in itself. Possible worlds are not static but dynamically time-dependent. Today’s world is not tomorrow’s world. This means that each possible world evolves in time according to rules. These dynamics can be represented by a tree diagram with nodes and branches reflecting the relations between the different worlds (nodes). Each branch is a retinue of possible worlds. A tree diagram reads top-down so that from one certain node, access is not given to any other node in the tree.

It is posited that A → B in the world m if and only if in all the worlds n accessible from m, A and B are simultaneously true. ¬A is true in the world m if and only if A is false in all the worlds n accessible from m.

Let us give examples of inference employing modal operators. Consider a situation S with two associated worlds S₁ and S₂ and two sentences a and b that can be true (T) or false (F) as shown in Figure 2.6.

Consider the inference It is invalid: a is T at S1; hence, ◊a is true in S. Similarly, b is true in S2; hence, ◊b is true in S. But the (a & b) is true in no associated world; hence, ◊(a & b) is not true in S.

By contrast, the inference is valid. If the premises are true at S, then a and b are true in all the worlds that are associated with S. Then, a & b is true in all those worlds, and □ (a & b) is true in S.

Each software user exposed to two environments is subject to developing a situation of his/her own in terms of rules of meaning and action and of sets of “rational” propositions pertaining to his/her situational world at a certain time. Software designers are not aware of this immanent state of affairs. But even if they were, they could hardly cope with the wide variety of possible situational worlds users of software systems are embedded in.

2.5.3.1. General considerations

Coordinating multiple agents is viewed as a planning problem. In this context, all actions and the interactions of agents are determined beforehand, leaving nothing to chance. There are two types of multi-agent planning, namely centralized and decentralized.

In centralized planning, the separate agents evolve their individual plans and then send them to a central supervisor which analyzes them and detects potential conflicts (Georgeff 1983). The idea behind this approach is that the central supervisor can:

• a) identify synchronization discrepancies between the plans of the stakeholders;
• b) suggest changes and insert them in a realistic common schedule after approval by the stakeholders.

The distributed planning technique foregoes the presence of a central supervisor. Instead, it is based on the dissemination of each other’s plans to all agents involved (Georgeff 1984, Corkill 1979). Agents exchange with each other until all conflicts are removed in order to produce individual plans coherent with others. This means that each stakeholder shares information with its partners about its resource capacities.

2.5.3.2. E-enabled coordination along a supply chain

To illustrate the role of Information Technology and especially telecommunications services in coordination planning, the example of e-enabled demand-driven supply chain management systems will be described (Briffaut 2015). By e-enabled, it is meant that all transactions are electronically paperless engineered along the goods flow from suppliers to clients. In particular, this context implies that customers place orders via a website. These cybercustomers expect to be provided without latency with data about product availability and delivery lead times.

The role of information sharing is acknowledged to be a key success factor in Supply Chain Management (SCM) in order to secure efficient coordination between all the activities of the stakeholders involved along the goods pipeline. Coordinating multi-agent systems (MAS) is realized in the case of an e-enabled SCM through a common information system engineered to share the relevant data between the stakeholders involved. The MAS approach is a relevant substitute for optimization tools and analytical resolution techniques whose efficiency is usually limited to local problems, without any adequate visibility over the behavior of the entire chain of stakeholders involved. Optimization of the operations of a whole is different from optimization of the operations of its parts A global point of view is required to bring under full control the synchronization of inter-related activities.

In traditional contexts, a front office and a back office can be identified in terms of interaction with customers. The front office carries out face-to-face dealings with customers while using a proprietary information system. Relationships with the other agents of the supply chain take place by means of messages exchanged between their information systems. Coordination between the front office and the back office is generally asynchronous and does not meet real-time requirements.

When e-commerce is implemented via a website, the delineation between the back office and the front office of the previous configuration is blurred and has no reason to be taken into consideration. The two offices merge into one entity because of the response time constraint. When queries are introduced by cybercustomers via a website portal, the collaborative information system must have the ability to produce real-time answers (availability, delivery date). Then, the information systems of the various stakeholders along the supply chain have to be interfaced in such a way that coordination between them takes place synchronously. The thing to do is to implement a workflow of a sort between the “public” parts of the stakeholders’ information systems. In other words, some parts of the stakeholders’ information systems contribute to producing a relevant answer to cybercustomers’ queries. Figure 2.7 shows the changes induced by a portal in terms of information exchange between the supply chain stakeholders.

2.5.3.3. Scenario setting for placing an order via a website

Each time a customer enters an order via the website, the supply chain collaborative information system shared by all stakeholders proceeds with an automatic check of projected inventories and uncommitted resource capacities per time period. If the item ordered is already available, the customer is advised accordingly and can immediately confirm the order. If the item is not available in inventory, the system checks whether the quantity and the delivery date requested can be met. In other words, the system checks whether manufacturing and supply capacities are available to meet the demand on the due date. Otherwise, it checks what the best possible date could be and/or what split quantities could be produced by using a simulation engine. Then, the customer is advised of the alternatives and can choose an option suitable to him/her. Once the option is confirmed, the system automatically creates a reservation of capacity and materials on the order promising system and forwards the order parameters to the back-office systems to be included in the production plan of the stakeholders involved. Order acknowledgments and confirmation are generated and sent by email.

2.5.3.4. Mapping the order scenario onto the structures of a BDI agent

The role of the Beliefs structure is to record order entries, send answers and process transaction data to turn them into memorized statistics. These statistics are used as entry data to update the APS (Advanced Planning System). The APS is implemented as a control tool over a short time horizon and is used as a non-repudiable commitment taken by the manufacturing shops. The Plan structure establishes and memorizes the APS pertaining to the supply chain as a whole. This means that this entity updates the supply chain APS on a regular time basis from data provided by the Beliefs structure. As the BDI agent acts as the front office of the supply chain with respect to the buyer side, it seems reasonable to ascribe it a centralized coordination role. Within this perspective, it draws up partial plans for the stakeholders on their behalf. When conflicts arise, it has the capabilities to bring the unbalance of distributed resources to terms. In other words, the Plan structure is assigned to implement the APS concept. It ensures that data required to derive their partial APS are made available in due time to all agents involved along the goods flow. It has two major features:

• – concurrent planning of all partners’ processes;
• – incremental planning capabilities.

APS is intended to secure a global optimization of all flows through the supply chain not only by increasing ROI (Return on Investment) and ROA (Return on Assets) but also by fulfilling customers’ satisfaction and retaining their loyalty.

The Desires structure is in charge of supporting the use of the ATP and CTP parameters. ATP stands for Available-To-Promise and CTP Capable-To-Promise. Per period of time, ATP has the capability to deliver an answer to a client request in terms of availability (quantity and delivery date). Either the request can be fulfilled from a scheduled inventory derived from the enforced APS or simulation of a sort has to be carried out through the CTP parameter to send an answer to the client. The CTP parameter takes account of lead times required to mobilize potentially available resources and allows a real-time answer to customer requests when necessary. It results, when activated, in the production of a new APS. The APS technique is generally supposed to be able to produce rolling manufacturing plans to match the demands of the buyer side.

The fulfillment of the committed schedule APS is ascribed to the Intentions structure. The PAB parameter is managed by this structure because it includes all of what is recorded as committed (APS and customer orders). The PAB (Projected Available Balance) parameter represents the number of completed items on hand at the end of each time period. It can be viewed as a means for giving some margin in cases of temporary unserviceable capability of some resources.

Let us use an example to explain how the roles of the four structures (Beliefs, Desires, Intentions and Plans) are connected and how they are performed by the interpreter. An input message coming from a customer is captured by the Beliefs data structure and analyzed by the Interpreter. It deals with a change in the delivery schedule induced by a new supply order entry. If this change requirement falls within the time fence linked to the supply lead time, it is rejected. Otherwise, as the agent has to act, the available resource capacities for the time period, be they committed or uncommitted, the projected on-hand inventory and the uncommitted planned manufacturing output are analyzed. If one of the possible supply sources can meet the specification required, it needs to select appropriate actions or procedures (Plans) to execute from a set of functions available to it and commit itself (Intentions). This simple scenario can be conceptualized by a repeat loop as shown below:

BDI-Interpreter

Initialize-state [ ];

repeat

• a) Options:= read the event queue (Beliefs) and go to option ATP (Desires);
• b) Selected option ATP = if ATP parameter proves relevant (order fulfilled without altering the existing MPS) then update its value and go to Intentions to update PAB , otherwise go to selected option CTP ;
• c) Selected option CTP = if CTP proves relevant (possible adjustment of current MPS while abiding by commitments in force) then go to Intentions for updating otherwise reject the request
• d) Execute [ ];

end repeat

2.8.3.1. The burgeoning backdrop

In the first era of the Internet, management thinkers talked up the networked enterprise, the Flat Corporation and open innovation as new business ecosystems, successors to the hierarchies of early-20th-Century industrial corporations. On one hand, these hierarchies remain pretty intact in big dot-com companies. On the other hand, new types of problems have emerged in terms of privacy, security and inclusion. They were solved by cryptography. New technologies, namely interconnected devices (Internet of Things), mass data storage, worldwide distributed ledgers (blockchain), etc., have enlarged these problems to such a scale that traditional organizational patterns in some economic sectors will be globally and locally overturned. It is not the purpose of this book to try to forecast how organizations will experience fundamental transformations in their structures and operations. An exercise of this sort is always hazardous. Let us focus on AI-driven autonomous agents which are coming into play. They can be defined as AI-driven devices that, on behalf of their designers, take information from their environments and are capable of choices, taking decision courses. They can modify the ways to achieve their objectives, sensing and responding to their environments over time.

Humans can interact with agents capable of varying degrees of autonomy whether in the loop (with a human constantly monitoring the operation and remaining in charge of the critical decisions), on the loop (with a human supervising the agent that can intervene at any stage of a process in progress) or out the loop (with the agent carrying out its mission without any human intervention once launched). In this last situation, the autonomous agent is under potential threats of cyber-attacks without being able to detect them and take appropriate counter-measures. Crucially, the identity of the attacker (another autonomous agent?) may be ambiguous, leaving those under attack uncertain as to how to respond when aware of the situation. This context is prone to change the very nature of the economic-competition battle field.

Any individual in our digital economy will more and more interact with faceless hidden partners, in most cases, without knowing whether the response received is sent by a human or a machine. Dealing with hidden partners is a source of uncertainty and anxiety leading to chaotic behaviors.

2.8.3.2. The looming artificial-intelligence-driven organization in the era of the Internet of Everything (IoE)

Organizations, whether they are populations, societies, groups or economic sectors, will have to come to terms with the co-presence of human agents with their psychological profiles in terms of beliefs, attitudes and knowledge and virtual agents, some of which are endowed with AI capabilities for deciding on action, communicating, reasoning, perceiving their environments, and planning their objectives.

The idea of DAC (Decentralized Autonomous Corporations) companies with no directors is being touted. They would follow a pre-programmed business model and would be managed by the applications of the block-chain tenet. In essence, the block-chain tenet means a shared mass agreement of transactions between a closed cluster of stakeholders and distributed storage of encrypted data. It is a ledger that keeps record of transactions accepted by all stakeholders and secures their storage. It is thought that the block-chain system would act as a way for a DAC to store financial accounts, insurance contracts, bonds or security records between the cluster members. This organizational architecture is appealing because outside intruders will find it hard to get access to encrypted data or to shut down the whole system in view of its decentralization. Airing these ideas may be considered as science fiction and disruptive with respect to the current context. But virtual agents have been in operation for a long time, for instance computer-supported Management Information Systems (MIS) and new types are showing on the economic stage. From finance (banking, payments, crowd-funding) to sharing economies (Uber and AirBnB-like platforms) to communications (social networks, email) to reputation systems (credit ratings, seller ratings on e-commerce sites) to governance, decentralized autonomous agents are already economic actors. Their possibilities seem endless in eliminating human intermediation in many industries, but trees do not grow up to reach the sky. Platforms like these may have massive implications on what the future will look like. When this current picture of the future face of the global economy is offered to the imagination of the public by futurologists, the basic characters of the laws of Nature should not be forgotten, namely nonlinearity, self-reorganization and chaos.

A special attention should be given to biology laws. Economic societies like all human and animal communities are composed of living organisms, and as such, reference to the approaches and concepts developed in the realm of biology could be helpful for understanding the evolution of economic societies. Jacques Monod’s seminal book Le hazard et la Nécessité – Essai sur la Philosophie Naturelle de la Biologie Moderne (Chance and Necessity – Essay on the Natural Philosophy of Modern Biology) (Monod 1971) is inspired by a quote attributed to Democritus “Everything existing in the universe is the fruit of chance and necessity”. Monod contends that mutations are unpredictable and that natural selection operates only upon the products of chance. In Chapter 7, Monod states that the decisive factor in natural selection is not the “struggle for life” but the differential rate of reproduction. The only mutations “acceptable” to an organism are those that “do not lessen the coherence of the teleonomic (end directed) apparatus, but rather, further strengthen it in their already assumed orientation” (Monod 1971, p. 119). Jacques Monod explains that the teleonomic performance is assessed through natural selection, and that this system retains only a very small fraction of mutations that will perfect and enrich the teleonomic apparatus. Jacques Monod makes the point that selection of a mutation is due to the environmental surroundings of the organism and the teleonomic performances.

What will be the balance of power between human agents and AI-aided virtual agents? It is likely that the future context will be an adapted evolution of the current one. What is impossible to forecast is the tempo of this evolution. In the 1980s, some futurologists forecast that by the end of this decade, cash money would have disappeared. Recently, the CEB has committed itself to keep printing bank notes. It is not a matter of technology but of psychology. People feel that they have the full control of cash money they can hoard. Let us try to elaborate on the position of human agents in a “mixed” context, drawing on the basis of existing situations.

Up to now, an important parameter has not been taken into consideration in the arena of Management Information System (MIS) design, namely the virtual nature of computer-aided information systems. A human making use of computer-aided information systems to collaborate in a business environment is exposed to a multiverse, a real-life universe and a virtual universe where interaction with a set of actors hidden behind a screen takes place. Figure 2.9 portrays this situation.

Each stakeholder of a collaborative computer-based context is at the interface of two environments with which they interact, namely a virtual environment via a human–machine interface and a real-life environment accessible through all their senses (sight, touch, hearing, smell) as shown in Figure 2.9.

When a decision-making process is engineered, this human relies on a space of pertinent facts extracted from the real-life and virtual universes they are exposed to. In order to conceptualize how this space is operated, the theory of the coordinated management meaning (CMM) can be called on. It is made up of schemata resulting from interpretative rules of meaning, deciphering messages and events coming from a real-life universe and from a virtual universe (computer-aided systems). Regulative rules of decision for action in the real-life universe are applied to derive data-driven decisions from memorized schemata. These rules refer to modal logic that we practice, often implicitly.

In any situation where human and virtual agents (actors) interact, a context is made up of three dimensions:

• – the agents involved characterized by their functional capabilities and roles;
• – the shared objectives and/or a common vision evolved from overlapping individual objectives;
• – the environment surrounding the cluster of agents bound and committed to reach a common achievement.

As it is stressed in second-order cybernetics, what is ascribed as the environment of a system, here a cluster of agents, must be considered as an agent of the system. The following arguments are intended to demonstrate it.

Message transmission between a sender and a recipient is prone to a distortion of content, often called filtering. This may incur biased interpretation by the receiver. It is an important feature to take into account when analyzing how agents communicate. It refers here to two notable elements. The first one is the semantics of messages and can be considered as a dimension of semantic interoperability. The other is the fact that messages can deliver a biased understanding of the actual context, engineered by the senders. A common basis of knowledge should be ensured between all the actors involved in order to avoid misinterpretation and, as a consequence, inefficient decision-making resulting from ambiguous mutual understanding. But accurate communication without the impairment of meaning is seldom, if ever, found in the complex realities of business life. Figure 2.10 portrays the picture of an interactive context.

Two interaction patterns between the actors, i.e. direct interaction and indirect interaction, can be identified. The idea of direct interaction is straightforward to understand. Indirect interaction means that the environment is a mediator that can influence the behaviors of the actors involved through the environment that is considered as a full actor of the system. This feature underlines the importance of the environment as an entity for context-awareness in collaborative environments. With this point of view, the environment must be considered not only as an actor in itself but also the only communication channel between the set of actors when a medium-based virtual collaborative environment is implemented. The behavior of each actor in the set cannot be understood and analyzed without taking into consideration how the interacting environment is perceived by each actor and is operative. Figure 2.9 changes into Figure 2.11.

A virtual collaborative environment is an artefact of a sort to which different members of a community are given access and that acts as an intermediate between them, either to exchange information and knowledge about a technical field of common practice or to help solve problems and to deliver results. In the current context, virtual collaborative environments are considered computer-based information systems shared by members of communities of practice, coming together virtually to exchange information and help members make pertinent decisions.

The universe each stakeholder is embedded in can be conceptualized as a space of tangible facts considered as relevant for making a decision in the real-life environment. To describe how this space is operated, the theory of the coordinated management of meaning (CMM) can be called on. It is the most comprehensive rule theory of communication developed by Pearce and Cronen (1980). CMM states that individuals in any social situation want to understand what is going on and apply rules to figure things out. In other words, constitutive rules in CMM are rules of meaning and rules of decision for action.

Rules of meaning are used to decipher a message or an event via interpretative rules. Rules of decision for action are regulative rules used to process interpreted messages or events. Rules of meaning and rules of decision for action are always context-dependent. Often, text (message or action) and context form a loop so that each is used to interpret the other (Cronen, Johnson, Lannamann) (Cronen 1982). The space of tangible facts for each business actor is a repository comprising interpreted messages from information systems and relevant events from the business actor’s real-life environment, events resulting or not from the business actor’s usual courses of action.

Let us delve into the background mechanism of the CMM rules of meaning. Kant uses the word “schema” when he argues, in his book, Critique of Pure Reason, that in order to apply non-empirical concepts to empirical facts, a mediating representation is necessary. He calls it a schema. In the ordinary case, there is, according to Kant, homogeneity of a certain sort between concept and object. There is no similar homogeneity in the application of concepts such as the intuitive analysis of messages and events. To apply a causal analysis to a sequence of messages and/or events so that one of them is regarded as a cause, another as the effect is more problematic, because the concept of causality involves necessity. But necessity is not always an element in our experience. The concept of causality only tells us that first, there is something and then something else. For example, the schema for causality is temporal succession according to a rule; the schema for necessity is the existence of an object at all times. Figure 2.12 shows the conceptualized processing steps for diagnosing messages and events.

Rules give a sense of which interpretations and decisions for action appear to be logical or appropriate in a given context. This sense is called logical force. But this logical force has a contextual dimension. The issue raised at this point is how to make fixed logic rules and changing contexts compatible: our answer is modal logic whose main features have been developed in a previous section.