A Multi-Agent Model of Workgroup Behaviour in an Enterprise using a Compositional Approach Meghendra Singh, Mayuri Duggirala, Harshal Hayatnagarkar, Vivek Balaraman Tata Research Development and Design Center 54-B, Hadapsar Industrial Estate, Pune, India +91-20-66086333 {meghendra.singh, mayuri.duggirala, h.hayatnagarkar2, vivek.balaraman}@tcs.com ABSTRACT The behavioural and social sciences domain have had a long and Fine grained behavioural models of humans in a context, such as rich tradition of human behavioural studies focusing on developing associates in an organization or a consumer group may allow us to theories of behaviour applicable across a wide range of settings both better understand existing behaviour as well as what would such as education, health, society and most significantly, happen given various situations. We have been working on a organizations. The methods used in behavioural and social sciences behaviour compositional approach to this problem which creates have focused on studying behavioural dimensions of interest using realistic agent behaviour models that when run as a simulation surveys, interviews and experiments with a relatively specific focus allows us to understand the dynamics of how various behavioural on identifying relations among the behavioural variables of interest. characteristics impact one or more outcome metrics of interest. The These studies use cross-sectional methods, i.e. study of behaviour compositional approach is founded on a repository of fine grained at a specific point in time, or repeated methods, that study behavioural relations and elements that have been mined from behaviour over a specific time period. However studies of dynamic ongoing and past research in the behavioural sciences as well as aspects of behaviour and the impact of behavioural dynamics on data obtained through field studies. The repository, the outcomes of interest are few and far between. Further, given the compositional system and the simulation system are part of an complexity, breadth and context sensitivity of human behaviour, it architecture we are developing called the Behaviour Analysis may even be infeasible or inappropriate to conduct experiments and Framework. We demonstrate our approach by showing how we can observational studies. In such constrained situations, in-silico create a behavioural model of specific aspects of a support team in simulations of human behaviour models become an attractive an enterprise. The goal of our model is to study how individual technique to explore, explain and perhaps predict how behaviour of behavioural dimensions such as emotion and stress impact the an individual unfolds over time in a given context. macro level outcome metrics of absenteeism and productivity. We Given the uniqueness and context-specificity of an individual’s use an agent based simulation to implement our behavioural model. actions it is computationally difficult to model all the possible We present some experimentation and results obtained from this motivations and other behavioural drivers of a given action. For simulation. Our simulation was able to reproduce various real- example, a person may respond to stress at work by either putting world phenomenon such as how absenteeism triggers more extra hours into the work or withdrawing from it completely. The absenteeism and slow recovery from workload spikes. drivers for this stress response could range from the person’s health status, skill level, reward systems, gender, supervisory support, CCS Concepts organizational culture, and many more such dimensions. •Computing methodologies ➝ Agent/discrete models Simulations of agent based models provide an approach to execute •Computing methodologies ➝ Multi-agent systems such computational models of human behaviour, such that at every simulation step, an individual agent makes a decision and takes an Keywords action considering its own internal behavioural state, behavioural Human Behavioural Modelling; Enterprise Modelling; Agent relations and context. In addition, in the presence of a large number Based Simulation of behavioural dimensions or variables, the analysis of their combined effect on outcomes of interest may turn out to be 1. INTRODUCTION intractable, inconclusive or even impossible. Hence the Organizations have been the focus of many research domains at computational model should find a way to generalize by accepting several levels of analyses-the macro or organizational level, the simplifications and approximations as useful compromises. team level and the micro or individual level. Computational modelling of organizations has traditionally focused on modelling This paper describes a framework for human behaviour modelling of macro-level organizational phenomena such as intra-and inter- and simulation in organizations called the Behaviour Analysis organizational networks, with less focus on realistic modelling of Framework (BAF). The BAF makes use of behavioural relations micro-level individual behaviour and its impact on team and extracted from past research and field studies in organizations to organizational level behavioural dimensions and outcomes. compose a human behavioural model. This model is then used to simulate an enterprise support services team. The simulation allows Copyright © 2016 for the individual papers by the papers' authors. us to perform experiments on team-level outcomes in the absence Copying permitted for private and academic purposes. This volume is and presence of the composed human behaviour. published and copyrighted by its editors. The rest of the paper is organized as follows: we present an overview of past research in human behaviour modelling and 2nd Modelling Symposium (ModSym 2016) - colocated with ISEC 2016, Goa, India, Feb 18, 2016 10 simulation, followed by our study approach, working example, In this work we discuss an approach to computationally model findings, insights and discussion on future work. human behaviour by combining results obtained from behavioural research literature and field studies. The approach utilizes the 2. PAST WORK Behaviour Analysis Framework (BAF) for collecting and analysing Human behaviour has been a topic of continued interest in a wide data, model creation/discovery and simulating a situation of variety of domains whether it is organizational behaviour or the interest. The particular example described here attempts to create a behaviour of consumers or citizens in society. The domains of realistic behavioural model of individuals in a software support and behavioural sciences, particularly psychology, sociology and maintenance team. We use behavioural insights from a field study anthropology have offered deep insights into human behaviour in a of a large support services organization, along with behavioural variety of contexts. . With respect to behavioural sciences, a variety models extracted from relevant literature to construct an agent of theoretical and empirical approaches are used to evolve new based simulation model of a support services team. hypotheses within a context as well as to test and validate them. Empirical techniques in the social sciences range from the We consider behavioural characteristics such as perceived quantitative, wherein surveys, experiments, meta-analysis, etc. are workload, affect and stress on the job; and the individual’s used, to the qualitative which involves an in depth examination of perceptions of her own daily productivity as the focus of the present text or qualitative data. Within the behavioural sciences, however, study. We refer to affect as the emotional experience of the the use of computational modelling in studying human behaviour individual at the start of the work day and the end of the work day, is relatively limited and is emerging as a technique to discover new with respect to their job. We use the positive affect negative affect theoretical relationships only in recent years. scale (PANAS) to measure various dimensions of positive and negative affect over the course of the workday for the respondent In recent years, there has been an emerging interest in the use of [8]. Stress at work has been measured using the model proposed by computational techniques to study human behaviour in various [11]. This model proposes distinct styles of decision making under contexts such as in an organizational setting [1]. The techniques stress, such as vigilance, hyper vigilance, and defensive avoidance. range from using apps to capture data to simulation engines to This model of stress and coping has earlier been implemented in simulate behaviour. The mix of computational techniques coupled simulators by Silverman et al [3]. with the social sciences provides a best of both worlds that can help in modelling and understanding human behaviour in various To summarize, previously developed approaches such as PMFServ domains of interest. Our paper uses this stream of literature to guide provide ways to model human behaviour in a performance-centric our overall approach in this study by the application of modelling manner. However, considering generality and modularity as techniques to understand behaviour in organizational settings at a essential elements to incrementally develop behavioural models, it more granular level. Given the relative novelty of the use of prompts us to have a fresh look at the problem and perhaps develop modelling and simulation in industrial and organizational a different approach, such as guided by software engineering psychology, we demonstrate how dynamic aspects of behaviour methods. can be examined in a specific organizational setting, like a software support services organization. 3. OUR APPROACH As discussed above, the primary focus of this work is to develop a With respect to the business context, support services are realistic model of human behaviour in the organizational context. significant drivers of growth and profitability for organizations Towards this goal, we have designed a computational framework today. Given the emphasis on business performance, the industry as shown in Figure 1. Behaviour Analysis Framework. The tracks a variety of metrics, most notably client-driven service level framework is called the Behaviour Analysis Framework (BAF) agreements (SLAs), process quality standards, other internal using which we collect field data, combine it with insights from organizational data and benchmarks for performance, which makes past research and use the combined behavioural relations in a it a rich context for an empirical investigation. Moreover, being a simulation. sector completely driven by its people, the support services industry, and more specifically the support services organization itself, offers a rich canvas for modelling various dimensions of human behaviour, associated task characteristics and their impact on work outcomes such as task completion, absenteeism and productivity. The quest of bringing the behaviour of simulated agents as close to the reality as possible draws in to account for various human factors such as personality, emotions, relationships with others, stress, goals, standards and preferences, and their impact on the agent’s decision making. Silverman et. al. [2,3,4] introduced a performance-centric approach namely 'Performance Moderating Functions', by choosing relevant models from behavioural sciences literature, to be later abstracted, composed and implemented for Figure 1. Behaviour Analysis Framework their computational platform 'PMFServ'. This platform is used as In the following sections we present our approach in more detail, a behavioural engine to drive synthetic agents in a military training in terms of data collection from the field, analysis, use of past simulator [5,6]. However, modelling, representation, simulation research insights and simulation. and prediction of human behaviour, as individuals and as in groups, still remains a challenge [7]. In addition, we believe that modularity 3.1 Mine Past Research and reusability of these models across various contexts and The core of the BAF is a behavioural relations repository. Relations domains adds a compounding difficulty. refer to models of behaviour which link specific behavioural 2nd Modelling Symposium (ModSym 2016) - colocated with ISEC 2016, Goa, India, Feb 18, 2016 11 dimensions to specific outcomes (e.g. the relation between affect to generate dynamic behaviour for a set of simulated entities and productivity) or between one behavioural dimension to another (agents) like, members of a support services team. The simulation (for example, a relation between Negative Affect and Stress). This engine executes the behavioural model for each agent given a is a store of behavioural relations extracted from past research and particular simulation scenario. Such simulations enables us to behavioural studies in the areas of behavioural and social sciences analyse the composed model’s behaviour in various scenarios, and that details the relationship among human behavioural dimensions reason about the closeness of the simulated behaviour to that and outcomes of interest. These dimensions could be aspects of an observed in the real world. Such a model can be used as a sandbox individual’s behaviours such as personality, affect, motivation, for discovering the impact of executive policies on important and performance, productivity, etc. The behaviour repository thus emergent outcomes like productivity of the work force, thereby offers us an extensive resource for behavioural insights to be used making it a useful decision support tool. for modelling and simulation. 4. Working Example 3.2 Collect Behavioural Data Behavioural data collection spans both quantitative (e.g. surveys, 4.1 Context and Goals secondary data, sensor data etc.) and qualitative methods (e.g. The present study was motivated by a request from a support interviews, focus groups, etc.). Duly anonymised, such multi- services organization requesting the research team for inputs on modal data sources can offer us an in-depth perspective of human behavioural drivers of absenteeism and productivity, which were behaviour in an organizational setting. the primary business priorities to their business leadership. The support services team wished to also explore the dynamics of the We start the study by exploring a specific organizational context, impact of various behavioural drivers on absenteeism and in this case a support services organization. We carry out in depth productivity. interviews with key stakeholders in the organization to get a deeper understanding of the business, its challenges, nature of work and its 4.2 Field Study and findings key deliverables. We also collect existing objective data in the We started the study by carrying out a preliminary field business context such as HR data on employee profiles, metrics on investigation in the support services organization mentioned above. performance, ratings, quality, productivity etc. Based on interview This enabled us to identify behavioural insights from a real business insights and a review of past research, we develop surveys aimed context, combined with other relevant objective data in the model at measuring dynamic aspects of behaviour such as affect, stress we developed. Further, we reviewed past research related to the and productivity. These surveys are administered at the start and individual, job, absenteeism, stress and productivity which revealed end of the workday to the respondents after seeking their informed specific behavioural relations. These were used in combination consent and after complying with due privacy and confidentiality with the insights for the field studies in our simulation model. We policies. discuss each of the specific aspects of this approach in the next 3.3 Analyse and Model section. Survey data are analysed using standard statistical techniques (e.g. 4.3 Collect Data regression, t-test, etc.). Survey data from the respondents and their We used a survey based approach for the field investigation where related organizational data help establish relations between our we measured various aspects of individual behaviour using behavioural dimensions (e.g. affect and stress) and outcomes of repeated surveys using an Android-based application on the user’s interest (e.g. absenteeism and productivity). smart phone over a two-week period. These surveys sought the Results from the survey data analysis coupled with a related search individual’s response on dimensions such as affect, stress and self- of past research from the behaviour repository, focused on the same reported productivity. behavioural dimensions, provide us with relations. As mentioned The survey was complemented by organizational data available above, relations refer to mathematical models of behaviour which from the HR teams of the support services organization. The link specific behavioural dimensions to specific outcomes. These organizational data provided us information on the individual’s relations are stored in the behavioural relations repository present demographic characteristics like location, education, skill level, in the BAF. tenure etc., as well as data on process quality, performance ratings and objective productivity. 3.4 Behaviour Composition and Simulation In order to realistically generate human behaviour in a simulated Together with the field study, we also carried out a review of entity (agent) we need a human behavioural model that ties literature, which demonstrated to us various behavioural relations behavioural elements like stress and affect to outcomes of interest that we examined in our study. The specific relations we studied like on the job productivity. We use the relations stored in the are shown in Table 1. behavioural relations repository of BAF to compose such human The surveys were administered to a sample of 100 volunteers from behavioural models for simulated entities. The composition begins the support services organization. Respondents were duly briefed with identifying common behavioural elements across relations in on the survey and their informed consent was sought before they the repository. These are then used to combine multiple relations participated in the study. The surveys were administered on the into one composite model of human behaviour for a simulated respondents’ Android smartphones and they were required to take entity. For example, the three relations, 𝑾𝒐𝒓𝒌𝒍𝒐𝒂𝒅 𝒑𝒆𝒓𝒄𝒆𝒑𝒕𝒊𝒐𝒏 → the surveys measuring affect, stress and productivity at the start and 𝑨𝒇𝒇𝒆𝒄𝒕; 𝑨𝒇𝒇𝒆𝒄𝒕 → 𝑺𝒕𝒓𝒆𝒔𝒔 and 𝑺𝒕𝒓𝒆𝒔𝒔 → 𝑷𝒓𝒐𝒅𝒖𝒄𝒕𝒊𝒗𝒊𝒕𝒚 are end of their workdays, over a two-week period. combined as: 𝑾𝒐𝒓𝒌𝒍𝒐𝒂𝒅 𝒑𝒆𝒓𝒄𝒆𝒑𝒕𝒊𝒐𝒏 → 𝑨𝒇𝒇𝒆𝒄𝒕 → 𝑺𝒕𝒓𝒆𝒔𝒔 → 𝑷𝒓𝒐𝒅𝒖𝒄𝒕𝒊𝒗𝒊𝒕𝒚 The composed model can be thought of as an aggregate of multiple relations present in the relations repository of the BAF. The composed model is then used by an agent based simulation engine 2nd Modelling Symposium (ModSym 2016) - colocated with ISEC 2016, Goa, India, Feb 18, 2016 12 Table 1. Variables of Interest Variable Data Sources Affect Survey Stress Survey Workload 1-1 interview Absenteeism Objective organizational data Figure 2. Composed model of human behaviour Productivity Survey The composed model figure 2 above takes the following form: The survey and objective organizational data were further supported by 1-1 discussions with delivery heads of sub-teams of 𝑾𝒐𝒓𝒌𝒍𝒐𝒂𝒅 𝒑𝒆𝒓𝒄𝒆𝒑𝒕𝒊𝒐𝒏 → 𝑨𝒇𝒇𝒆𝒄𝒕 → 𝑺𝒕𝒓𝒆𝒔𝒔 → 𝑷𝒓𝒐𝒅𝒖𝒄𝒕𝒊𝒗𝒊𝒕𝒚 the larger sample. This discussion helped us get information on 𝑾𝒐𝒓𝒌𝒍𝒐𝒂𝒅 𝒑𝒆𝒓𝒄𝒆𝒑𝒕𝒊𝒐𝒏 → 𝑨𝒇𝒇𝒆𝒄𝒕 → 𝑺𝒕𝒓𝒆𝒔𝒔 → 𝑷(𝑨𝒃𝒔𝒆𝒏𝒕𝒆𝒆𝒊𝒔𝒎) various process level metrics such as workload, hich has been used in the simulation model. Here, the perceived workload (Workload perception) is defined as the number of extra tasks arriving at the beginning of a work day 4.4 Simulated Model over and above the mean of task arrival for an individual agent. We created a multi agent simulation of a software support team Affect refers to a general positive or negative emotional experience using the approach discussed above. We used the GIS and Agent- of the individual at the start and end of the work day. Stress refers based Modelling Architecture GAMA [9] as a simulation engine to to negative behavioural responses to a specific challenging implement the simulation. The simulation model has two key situation or context. In our study, we refer to absenteeism as the components: 1) A process model that describes team and task- number of days the individual goes on unplanned or unscheduled related metrics for the support services team being simulated, 2) A leave from work. Lastly, productivity refers to the number of human behavioural model composed from relations present in the assigned tasks completed at the end of the person’s work day or BAF that establishes relationships between a simulated agent’s shift. Table 2 describes the specific equation-based or rule based behavioural characteristics and the functional outcomes of interest forms of relations used in the composed human behavioural model like productivity and absenteeism probability. We describe each of for the simulation: these components below. Table 2. Relations used in the composed human behavioural 4.4.1 Process Model model for simulation The simulated team consists of 50 individuals working in a typical support services workspace. The maximum number of work hours Relation Model Source allowed per agent is 10. This includes 8 regular work hours, along Negative Affect=0.53 Illies et 𝐴𝑓𝑓𝑒𝑐𝑡 ⃪ 𝑊𝑜𝑟𝑘𝑙𝑜𝑎𝑑 with a maximum of 2 hours of overtime. Work is modelled as (workload)+0.7 al.[10] discrete and independent tasks arriving at the beginning of each work day. The exact number of tasks arriving on a work day is 𝑆𝑡𝑟𝑒𝑠𝑠 ⃪ 𝐴𝑓𝑓𝑒𝑐𝑡 Field Stress=0.39 (Affect)+0.51 taken from a Normal distribution with a mean and standard Study deviation of 1000 and 100 respectively. These tasks accumulate in a task pool for the entire team. Individual tasks from this task pool If (Stress <= 0.1) then are then uniformly allocated to all the available members (agents) Productivity= 0.1 * Productivity of the support team. An available member is a team member (agent) If (Stress > 0.1 && <= 0.25) then Productivity= 0.5 * that is present for work on a particular work day. The initial Janis Productivity productivity per agent is set to 2.5 tasks per hour. A task once Mann If (Stress > 0.25 && <= 0.75) finished is removed from the task pool. 𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑣𝑖𝑡𝑦 ⃪ 𝑆𝑡𝑟𝑒𝑠𝑠 (1977) then Productivity= 1.25 * Silverma Productivity 4.4.2 Human-like Behaviour Model n’s PMF If (Stress > 0.75 && <= 0.9) The following relations were discovered after the analysis of the Model then Productivity= 0.5 * data collected from the field study and mining of relevant literature: Productivity If (Stress > 0.9) then 𝑾𝒐𝒓𝒌𝒍𝒐𝒂𝒅 𝒑𝒆𝒓𝒄𝒆𝒑𝒕𝒊𝒐𝒏 → 𝑨𝒇𝒇𝒆𝒄𝒕 Productivity= 0.1 * Productivity 𝑨𝒇𝒇𝒆𝒄𝒕 → 𝑺𝒕𝒓𝒆𝒔𝒔 𝑃(𝐴𝑏𝑠𝑒𝑛𝑡𝑒𝑒𝑖𝑠𝑚) ⃪ 𝑆𝑡𝑟𝑒𝑠𝑠 Field If stress>0.9 then N(0.1, 0.1) 𝑺𝒕𝒓𝒆𝒔𝒔 → 𝑷𝒓𝒐𝒅𝒖𝒄𝒕𝒊𝒗𝒊𝒕𝒚 Study 𝑺𝒕𝒓𝒆𝒔𝒔 → 𝑷(𝑨𝒃𝒔𝒆𝒏𝒕𝒆𝒆𝒊𝒔𝒎) The composed human behavioural model discussed above is used As discussed in section 3.4, we combine these relations together to by the simulation engine to generate human-like behaviour for come up with a composed human behavioural model. We began by simulated team members of a support services team constrained by identifying common behavioural variables i.e.: Affect, Stress and the process model discussed in section 4.4.1. Although the model used these to combine the four relations above into a composite may seem bounded by fewer parameters and finite value range, it behavioural model. The relations used in this paper thus have been generates interesting dynamics to study spectrum of outcomes. A tested in past research as well as validated in a field study within a model with large set of parameters and wide range of values real organization. This model takes into account two behavioural variables: Stress and Affect and two outcomes of interest: Productivity and Absenteeism probability. 2nd Modelling Symposium (ModSym 2016) - colocated with ISEC 2016, Goa, India, Feb 18, 2016 13 becomes intractable for analysis. Thus this can be considered as a axis represents the average level of stress in the team. This ranges boxed experiment. from 0 to 1. The simulation runs for 120 working days (24 work-weeks). A work-week is composed of 5 working days. A spike doubles the number of tasks arriving on a particular work day. 5. FINDINGS We run two scenarios on the simulation model discussed above. First, we execute the model without taking into consideration the impact of any human behavioural variable on agent behaviour. In the second case, we factor in the impact of two human behavioural variables: Affect and Stress on agent productivity and make observations on outcomes like the state of the task pool for the team and average stress levels for the team. 5.1 Agents as automatons Figure 4. Variations in the team’s common task pool In the first scenario agent’s behave like automatons working on an over simulation time assembly line. Here, if the number of tasks arriving at the beginning of a day are equal to or less than the combined daily work capacity of all the agents, the task pool is empty by the end of the work day. However, if the number of tasks arriving at the start of a day are more than the combined work capacity of all the agents, the task pool retains some tasks by the end of the workday, which results in task build-up or backlog. Consequently, a spike in task arrival leads to the formation of a substantial backlog. However, the system is able to recover from the backlog because of a work spike in a few days because of buffer staff and regular staff working overtime. Buffer staff refers to extra workers that can support the team operations at a time of crisis or work spike. The amount of buffer staff used in the simulation was 6% of the total staff. Figure 3 shows the task pool for the entire team over simulation time. Here, the Figure 5. Variations in the team’s average stress level horizontal axis represents simulated work hours and the vertical over simulation time axis represents the number of tasks present in the task pool. Based on the human behavioural model discussed earlier, it is important to highlight that an individual performs optimally only when her stress level lies in the green band of the graph (0.25 to 0.75). If stress for an individual moves out of this band, the individual starts performing sub-optimally leading to low levels of productivity, which leads to backlog and an accumulation of tasks in the team’s common task pool. We observe that the average stress in the team spikes as the workload on individual agents increases because of a spike in tasks, on the second week (after 50 hours on the time axis). However, as the workload stabilizes the average stress level returns to normal. The spike in stress also increases the absenteeism probability of individual agents, causing a spike in the number of absentees on the subsequent days of the spike in average stress. Figure 6 shows the Figure 3. Variations in the team’s common task pool variation in the number of people who are absent with respect to over simulation time the simulated work days. Here, we observe that absenteeism triggers a vicious cycle, leading to more absenteeism afterwards. 5.2 Agents with human-like behaviour In the next experiment we factor in human behaviour for each agent. We implement workload perception, affect and stress as behavioural variables for each agent, and implement the human behavioural model as discussed in section 4.4.2. We execute this experiment for the simulation duration of 120 working days and record metrics like: the state of the task pool over time, average stress among the individuals in the support team and number of absentees on a work day. Figures 4 and 5 show the task pool for the entire team over time and average levels of stress for the team over time respectively. For figure 4, the horizontal axis represents simulated work hours and the vertical axis represents the number of tasks present in the task pool. Similarly, for figure 5, the Figure 6. Variations in the number of absentees in horizontal axis represents simulated work hours and the vertical the team over the simulation days 2nd Modelling Symposium (ModSym 2016) - colocated with ISEC 2016, Goa, India, Feb 18, 2016 14 This happens because the absence of a team member causes the [3] Silverman, B. G., Johns, M., Cornwell, J., & O'Brien, K. tasks allocated to the team member to be distributed equally among (2006). Human behavior models for agents in simulators and the rest of the available team members. This causes an increase in games: part I: enabling science with PMFserv. Presence: the workload of the rest of the team leading to an increased stress Teleoperators and Virtual Environments, 15(2), 139-162. among the rest of the team members. Increased stress translates into [4] Silverman, B. G., Bharathy, G., O'Brien, K., & Cornwell, J. a greater probability of being absent on the subsequent day for the (2006). Human behavior models for agents in simulators and stressed team members. This leads to a greater number of team games: part II: gamebot engineering with PMFserv. members being absent on the subsequent days. Presence: Teleoperators and Virtual Environments, 15(2), 163-185. 6. FUTURE WORK AND CONCLUSION In this work we propose a model that incorporates human behaviour [5] Silverman, B. G., Pietrocola, D., Weyer, N., Weaver, R., into simulated agent’s working in an enterprise. We were able to Esomar, N., Might, R., & Chandrasekaran, D. (2009). factor in various elements of realism like backlog, absenteeism and Nonkin village: An embeddable training game generator for spikes in workload. The intent of this work is to demonstrate a learning cultural terrain and sustainable counter-insurgent possible approach of modelling human behaviour into programmed operations. In Agents for Games and Simulations (pp. 135- agents. The human behaviour model presented here is limited when 154). Springer Berlin Heidelberg. compared to real-world human behaviour in that it includes only [6] Silverman, B. G., Pietrocola, D., Nye, B., Weyer, N., Osin, bivariate relations. However, this approach allows the modeller to O., Johnson, D., & Weaver, R. (2012). Rich socio-cognitive begin with simpler models and then incrementally add more agents for immersive training environments: case of NonKin complex behaviour until sufficiently realistic behaviour emerges. Village. Autonomous Agents and Multi-Agent Systems, The model can then be used to understand a situation and devise 24(2), 312-343. interventions if necessary. One dimension for augmenting the current model is to factor in the social environment and social [7] Taylor, S. J., Khan, A., Morse, K. L., Tolk, A., Yilmaz, L., & influences. A team would thus be modelled not just as a set of Zander, J. (2013, April). Grand challenges on the theory of individuals but a network of networks of social and professional modeling and simulation. In Proceedings of the Symposium relationships. A related area of work which is notably difficult [12] on Theory of Modeling & Simulation-DEVS Integrative is to extract insights from data generated by simulation of complex M&S Symposium (p. 34). Society for Computer Simulation networked models with many behavioural variables. International. [8] Watson, D., & Clark, L. A. (1999). The PANAS-X: Manual We would also like to increase the number of behavioural variables for the positive and negative affect schedule-expanded form. that are at play. In particular, we would like to use various personality traits and their interplay with agent productivity. This [9] Amouroux, E., Chu, T. Q., Boucher, A., & Drogoul, A. would allow the modeller to compose teams of complex agents (2009). GAMA: an environment for implementing and having diverse personality traits, which would lead to different running spatially explicit multi-agent simulations. In Agent behaviours for the organization as a whole. We would also like to computing and multi-agent systems (pp. 359-371). Springer model more outcomes of interest, like attrition and change in agent Berlin Heidelberg. expertise. [10] Ilies, R., Huth, M., Ryan, A. M., & Dimotakis, N. (2015). Explaining the Links Between Workload, Distress, and 7. ACKNOWLEDGMENTS Work–Family Conflict Among School Employees: Physical, We are grateful to our group members Sachin Patel and David Cognitive, and Emotional Fatigue. Clarance for their contributions to this work. [11] Janis, I. L., & Mann, L. (1977). Decision making: A 8. REFERENCES psychological analysis of conflict, choice, and commitment. [1] 1. Weinhardt, J. M., & Vancouver, J. B. (2012). Free Press. Computational models and organizational psychology: [12] Ju-Sung Lee et al., (2015). The Complexities of Agent-Based Opportunities abound. Organizational Psychology Review, Modeling Output Analysis. In Journal of Artificial Societies 2(4), 267-292. and Social Simulation 18 (4) 4. DOI: 10.18564/jasss.2897 [2] Silverman, B. G. (2004). Toward realism in human performance simulation. The science and simulation of human performance, 5, 469-498. 2nd Modelling Symposium (ModSym 2016) - colocated with ISEC 2016, Goa, India, Feb 18, 2016 15