The New Frontier - The Role of Ambient Computing in Digital Representative Negotiation
Authors(s) - Wei Kong, Kevin Dragomir, Filipe Minho, and Calvin Lim
Saint Mary’s College of California, United States
School of Economics & Business Administration
The advent of technology and its subsequent growth into becoming a key component in contemporary society has been cited by scholars to have come with it numerous benefits. Top on the list of merits has been the introduction of the internet of things that uses sensors placed on a network of physical objects for the purpose of collecting and exchanging data and information. The objects used in the internet of things move further to analyze the collected and provide top tier actionable intelligence on the next course of action.
Bala argues that data analytics improves the quality of the data provided, which in turn meets the needs of individuals and groups, as shown in figure 1 below. This essay analyzes the application of the internet of things, specifically ambient computing, in the collection of data on the temperature in a room and the provision of actionable intelligence, which regulates thermal comfort to the set needs of the users. The research operates under the auspice of the thesis that the application of ambient computing can help in the mitigation of any dissonance between people when negotiating the right temperature level.
Figure 1: Application of Internet of Things in Modern Settings
Source: (I-Scoop)
Human interaction is largely transactional by nature with individuals and groups implementing different strategies that can help them attain their set goals and objectives. Most of these interactions are, however, compounded by personal preferences that influence the way the individuals think (Maiwald). This leads to increased dissonance and conflicts during transactional interactions as every person is trying to get something out of it.
Maiwald highlights that the introduction of negotiation in a conversation can go a long way in mitigating conflicts. The author defines negotiation as a process in which an agreement or a compromise is reached while avoiding any dispute and agreements. In addition, the negotiation aims at creating a conducive environment in which the parties involved get a beneficial outcome that addresses either part of their need or even all of it (Maiwald).
Maiwald goes further to indicate that negotiations require one to put forward a position on the issue at hand to gain an insight into what the other person thinks. The provision of a perspective is followed by the making of concessions with the parties trusting each other when it comes down to the implementation of the negotiated solution (Maiwald). The next step in the negotiation is for the parties to find opportunities that are inconsistent with theirs, which informs them on the type of negotiation the other person wants to take. They then move forth towards face-saving as they justify their stance on the matter (Maiwald). The identified steps are all dependent on the techniques used with active listening providing an individual with the necessary intelligence on the next step they need to take (Maiwald).
There are two major types of negotiation, which include integrative and distributive negotiation. With integrative coordination, the end outcome is predicated on the perspective of the two parties (Maiwald). It takes advantage of their differing world views with an individual losing something in exchange for another. Also, it constructs and reframes the issues of the conflict to ensure that the outcome is a win-win (Maiwald).
Distributive negotiation, on the other hand, is said to be a positional form that focuses on the unequal distribution of the benefits. Maiwald highlights that the distributive negotiation is dependent on the zero-sum theory, whereby one party will gain the benefit at the expense of the other. In most situations, the distributive negotiation is said to be a win-lose with the conflict exacerbating at the very end (Maiwald). One has to note also that the distributive negotiation normally occurs in instances where the parties have never had an interactive relationship before with both adopting a fixed position until the other cedes theirs (Maiwald).
There are numerous examples of people wasting time negotiating with one of them being the identification of a solution for one-size-fits-all room temperature. Assume it is on a sunny day, and there is a pregnant woman in the room with her husband. The woman states that she wants the husband to switch on the AC for air regulation, but he hesitates stating that the temperature is okay. They try to negotiate using distributive negotiation with the woman getting the upper hand because of her condition.
The second example is, say, a neighbor is blasting loud music in the middle of the night. One of the tenants approaches him and asks for the music to be shut down. The neighbor states that he is having a shindig, and turning off the music will not entirely cut it. Using the integrative negotiation technique, both parties agree that the party will limit the noise with the tenant holding out on calling the police. Both parties benefit from the negotiation, thereby limiting any conflict and dissonance.
Conflicts and problematic situations constitute a normal part of life. For example, in a working environment where various people interact, many instances in which different people or groups disagree may be witnessed. This is particularly fostered by the natural variations in people which cause differences in preferences, ideologies, and opinions, and so forth. As a result, the need to find solutions for such problems cannot be overstated. For instance, to mitigate the conflicts, people are likely to engage in a negotiation through which a solution may be obtained.
However, the process through which a solution is obtained may as well be problematic in itself. This is true, especially because of challenges associated with efficiency, effectiveness, and costs associated with the utilization of resources. As such, because problems and challenges are inevitable, it becomes important to devise optimal resolutions. For a given problem, there are various decision variables and several possible solutions based on how one chooses to exploit those decision variables. According to (Murty), the careful evaluation of the various solution alternatives to adopt the best possible solution is referred to as optimization.
An organization or team must have objectives and goals. In that light, every member of the team can be regarded as a functional resource. To achieve the desired effects of the team, every member of the team must not be distracted or disoriented towards his or her goal. Optimization is the process through which such distractions are erased and strengths are promoted instead (Murty). As such, it becomes necessary to devise optimization models with which actions can be easily governed. Murty further states that since ancient times, people have always endeavored to optimize their systems, and, via identifying objective functions to optimize specifically to manipulate decision variables for better results, people have continued to devise optimization models.
Concerning an operation, therefore, whether it’s a conflict resolution process or a milestone pursuit, performance becomes the basic measure of optimization. And, optimization follows either of two optimization model structures based on the number of objective functions. An objective function is a measurable variable or resource that can either be maximized or minimized as desired (Murty). Such would include time among other things. For example, in the case of a conflict, it would be most desirable to use the minimum time possible to reach a lasting and agreeable solution. Therefore, an optimization model that focuses on minimizing the time used up in resolving conflicts would be most suitable during negotiations.
The two optimization models are the single objective model and the multi-objective model. As the name suggests, the single objective model focuses on manipulating a single value. For example, however, a negotiation ends; optimization might be interested in minimizing the time used. In such a case, the model becomes a single objective. Multi-objective, on the other hand, aims at manipulating more than one variable. For instance, while the goal of a team may be to create a certain product, the optimization model may focus on minimizing the time used as well as maximizing the product’s quality. In such a case, the model becomes multi-objective.
Optimization, however, is not straight forward with multi-objective models. Perhaps it’s because of the challenges associated with priority and determination of performance measures. Nonetheless, to successfully optimize an operation, it is necessary to collect and analyze the relevant data and information. As such human/ machine learning technologies may come in handy. For example, to optimize a temperature regulation system, it becomes necessary to collect real-time data about the environment in question. That can be made possible via the use of human/ machine learning technologies.
As opined in the introductory tenet, the main purpose of the internet of things is the management of the divergent needs of the population with the time and costs incurred being limited. Ambient computing is defined as the combination of human/ machine learning and interaction together with user experience, software, and hardware. The application of ambient computing strategies can be beneficial in that it brings both the physical world and the digital world together. Walcott states that the digital and physical worlds merge as the intelligent cloud is becoming more consistent with the personal lives of individuals.
By using emotion, hearing, and vision on top of other cognitive services, ambient computing works towards understanding and managing the different needs of a populace (Walcott). Furthermore, the systems are cited as being invisible as they are embedded in the computing technologies around the individuals (Walcott). This will lead to less visual clusters and improved capabilities.
The second benefit of ambient computing is that it will improve socialization between humans and machines (Walcott). Smart buildings will be able to gain information on the occupants and elicit a social response that is tailored to the needs of the occupants. The third benefit is that it will improve on decision making by negotiating the different needs and preferences of the users through the use of special algorithms (Walcott). The smart environment will make the choices for the individuals, thereby minimizing any conflicts that might arise out of integrative and distributive negotiation.
The fourth benefit is that the systems will help improve the adaptability of buildings as they will yield proactive behavior rather than being passive (Walcott). Scholars argue that buildings will gain information that will allow them to learn user preferences and run both aesthetically and efficiently. The final benefit of ambient computing is that it will allow for and improve the levels of convergence (Walcott). Summations from the research show that much of the environment will be supplemented with the introduction of new digital technologies (Walcott). The network created will improve the manner through which information is shared, thereby eliminating mundane tasks. Furthermore, there will be few errors occurring in the information process with systems pulling data from shared digital locations.
To better understand the efficacy rates of ambient computing in real life, the researchers evaluate a case study in which three people in a room have three different preferences with regards to the temperature. One of them likes it hot, the other likes it cold, while the last one does not care. The dilemma in the room is whether the individuals need to turn on the AC, open a window, or turn on the heat. To meet the needs of all the participants in the room, the system uses behavioral science and social choice theory.
Science highlights that thermal comfort, for any organism, occurs when and only if the mind expresses satisfaction in the environment it is in (Guszcza). The systems thus conduct a subjective evaluation with the body being the reference point. The system analyzes the manner through which the heat from the body is transferred and whether it is proportional to the difference in temperature in the room (Bala).
The system also takes into account whether the body is losing heat or if it is not exerting any heat. This is conducted through the use of the predicted mean vote, which evaluates the heat load through behavioral sciences (Anika Schumann). The mean vote utilizes six different variables, with the first being humidity, relative air velocity, the radiant temperature, and the indoor air temperature (Bala). It also analyzes the clothing worn by the users and their activity levels (Bala; Anika Schumann). The PMV has to work with the ambient technology to ensure that at least 80% of the people in the room are satisfied with the thermal comfort.
Addendum to this, the ambient technology uses behavioral science to evaluate how the individual reacts when exposed to outdoor climate and how the latter influences thermal comfort. This information is significant in evaluating the divergent adaptation mechanisms used by humans in different temperatures during different seasons (Anika Schumann). The systems adopt past thermal history and integrative negotiation to influence the thermal expectations and preferences of the occupants.
Social theory can also be applied in the systems in that it evaluates the preferences of people and their satisfaction rates through the use of the percentage of people dissatisfied indices. The social theory is vital for the ambient systems as it makes generalizations and distinctions among the different types of people in a room (Guszcza). Scholars argue that the placement of sensors in rooms can help researchers define the degrees of similarity, after which the relevance functions are weighted with the situation (Guszcza). The systems take into account the preferences that supersede the other and implement them effectively. Furthermore, ambient technologies predict the comfort votes based on the data provided with both integrative and distributive negotiation being applied to ensure that the disadvantaged party gets a chunk of the benefits (Guszcza).
In our research, we use decision tree because most of our variable at classifiers such as season, climate type, room type. The decision is used to mimic an individual’s thought or decision process to create a “digital twin”. This is done because we needed a test twin to serve as a proxy for the ambient twin. In the ambient environment, a data twin will be provided. Our data is fairly limited and while the dataset size is large enough, the types of variable recorded are qualitative rather than quantitative, thus making decision tree the favorable option rather than Deep Neural Networks. In our model, 2 decision tree models are utilized with separate models using PPD (Predicted Percentage Dissatisfied) and preferred air temperature as the dependent variables. The decision tree was used with different goals and in different situations. In method one we use Decision tree to generate a function that would express the relation between PPD and Temperature for each individual. In order to do that, we first fit (built) the model by using EB (environmental bin) and Temperature from the whole cleaned data set to calculate the PPDi. Secondly, using the same model and the Simulation dataset we were able to predict the PPDi.
In method two, we generated thirty DT (Decision Tree) to present the relation between PPD and Temperature. To find the predicted PPDi, we use EB and Temperature from a subset dataset which corresponds to each individual. Later we generated another thirty DT, but this time to present the relation between Temperature and PPD. To find the predicted Temperature for each person, we built the models based on the sunset dataset from each person. We also use some theory of behavioral science in the optimization model with the "For Loop" method. The second objective function in this method reflects the idea that each person's preference is weighted before optimizing them in a group decision.
One of the main objectives of Ambient Computing is returning the privacy rights to an individual. In a tech oriented world, data privacy has become a prevalent issue. Using ambient computing and blockchain, we hope to keep personal preferences private. Using blockchain, small amounts of information will be transmitted on a need to know basis. Because of the unique features of blockchain and its ability to provide anonymity, combining this technology with ambient computing provides an intriguing method for privacy protection.
Works Cited
Anika Schumann, Nic Wilson, and Mateo Burillo. " Learning User Preferences to Maximise Occupant Comfort in Office Buildings." Trends in Applied Intelligent Systems (2010): 681-690.
Bala, Mohammad Irfan. "A model to incorporate automated negotiation in IoT." 2017.
Guszcza, Jim. The last-mile problem: How data science and behavioral science can work together. 27 January 2015. 11 November 2019 <https://www2.deloitte.com/us/en/insights/deloitte-review/issue-16/behavioral-economics-predictive-analytics.html>.
I-Scoop. The Internet of Things (IoT) – essential IoT business guide. 11 November 2019. 11 November 2019 <https://www.i-scoop.eu/internet-of-things-guide/>.
Maiwald, André. "Power, Negotiation Type, and Negotiation Types." 2015.
Walcott, Jeff Fedders & Katalin. Ambient World, "I Am Aware." 2018. 11 November 2019 <https://ieeetv.ieee.org/mobile/video/keynote-ambient-world-i-am-aware-jeff-fedders-and-katalin-walcott-ttm-2018-2?rf=events%7C158>.
Murty, Katta G. "Optimization models for decision making: Volume." University of Michigan, Ann Arbor (2003).