An integrated use of tools to acquire knowledge elements and support decisions in innovations processes

Working with socio-technical systems, where technological components are inter-related with the complexity that is generated by individual and organizational actions and processes, presents several methodological problems. When an innovation process develops in a socio-technical system, many of the involved factors are not meaningfully quantifiable, since they are connected to technological but also social, organizational, political and cognitive dimensions. Everything is connected to everything else and “what might seem to be the most marginal of factors can, under the right circumstances, become a dominating force of change“ (Ritchey, 2006). These innovation processes are characterized by multiple actors and perspectives, competitive or conflicting interests, constraints and uncertainties that (using the distinction proposed in Friend, 1989) can be connected to the working environment, the related decision fields and/or the guiding values. All these elements define what Rosenhead and Mingers (2001) called an unstructured problem. “How can you deal with the main complexity and uncertainty elements of an innovation process where multiple actors are or should be involved?”, but also “How could be facilitated decision in such problematic situations?”. Acquiring and organizing knowledge and information elements can be essential not only to understand but also to eliminate, reduce or control complexity and uncertainty. “Traditional quantitative methods, mathematical (functional) modeling and simulation will simply not suffice in several cases” (Ritchey, 2006). The sociological and psychological literatures suggest approaches and methodological tools to identify complexity and uncertainty and cope with. Logical and structured procedures are also proposed in the Operations Research (OR) literature (but also in fields such as System Engineering, System Analysis or Cybernetics) as “soft OR methods or problem structuring methods (PSM) ”, to facilitate a shared vision of the situation and to decide how complexities and uncertainties have to be controlled and improvement actions to be elaborated, evaluated, validated and implemented. The questions become “Which tools can be used to facilitate decision in such problematic situations?” and above all ”Can an integrated use of soft and traditional methods be more effective in these situations?” Problem structuring methods could be improved by an integrated and interdisciplinary approach that systematically helps in identifying or constructing an agreed framework for the problem formulation, above all when the situation is “new”. A trans-disciplinary approach is required with methods which can bring together social psychology, psychology, math, strategic management, logic and computer science (Eden and Ackermann, 2006). The aim of this research is to propose and test the integration of soft and traditional tools, that come from different disciplines, in order to propose a PSM, to identify, try to reduce and control uncertainties and complexities elements in innovation processes. In relation to this aim, approaches and tools, from social psychology and cognitive psychology, and the main concepts of the theory that is at the base of the proposed approaches are analyzed in the first chapters of the thesis. All the analyzed tools are classified in terms of reference theories, aim, source of the structured elements, adopted structure and content. In the first chapter of the thesis the relations between cognitive approach, cognitive psychology thinking, psychological theories and mapping techniques are analyzed, in order to acquire, synthesize, code and communicate all the elements that come from the different points of view of the actors who are or can be involved in a process that can be cognitive but also a decision process or an innovation process. The chapter is oriented to analyse why and how cognitive scientists study similarities between thinking and information processing, why people use some types of knowledge and discard others, built some types of map to represent knowledge and not others, why some people want to identify and define the knowledge flow in process or in what way it is possible to support the creation of knowledge in different stages of process in which the human thinking is centred. The analysis stars from the study of Neisser, Piaget, Kelly and Tolman, cognitive psychologists who want to understand how human “make sense of “ their world and try to manage and control the context around us. Different types of maps, that are proposed in the literature, can be created and modified by conscious intent or without conscious intent, showing cognitive structures, reflecting values, emotions, behaviours. The proposed maps were analyzed and classified, in order to understand how they can be elaborated and used in real processes. Conceptual maps (Novak, 1970) explore individual knowledge, represent and communicate expert knowledge and create new knowledge. Casual maps (Eden, 1992) are useful in strategic management for organizations, in order to diagnose the reasons for unsatisfactory and satisfactory outcomes, at the end of a project, to facilitate learning and risk analysis for future projects or help actors to identify potential difficulties and policies and reduce those difficulties, at the beginning of a project. Argument maps (Toulmin, 1958 and Horn, 1998) could be used in business settings to support the analysis of pros and cons. Mind and semantic maps (Buzan, 1996) are used to generate, visualize, structure and classify ideas, explain concepts behind words, associations and diachronic purposes and to predict language change. Knowledge maps (Rogers, 2000 and Stanford, 2000) visualize knowledge beyond textual for the purpose to elicit, code, share, use and expand knowledge. Topic maps (Newcomb, 2007 and Pepper, 2009) are standard representations and interchange of knowledge, with an emphasis on the information findability (how can be found information that is contained on a website). In the second chapter one mapping techniques, cognitive mapping, is deeply analyzed. Cognitive mapping aims to provide a tool for revealing subjective beliefs in a meaningful way so that they can be examined not only by the individual for whom the map is constructed, but also by other individuals and groups (Eden, 1992). Cognitive mapping is a technique for knowledge elicitation and recording (from an individual or during a discussion in a group) that facilitates the participation of multiple actors in a problem-structuring process, the emergence of a shared representation of the problem (starting from individuals’ views of a problem situation) and the elaboration of a problem-solving process that includes all the members of a group (Rosenhead and Mingers, 2004). A cognitive map has some structural properties (Eden, 2004) and the analysis can be centred on the whole structure of the map (shape, layout), on the nodes - concepts (central concepts, clusters) or on the relationships between concepts (loop, contradiction). Some methodologies of cognitive mapping were analyzed and some applications studied in order to understand their potentialities. The first methodology was originated from research into methods of helping the process of problem solving in teams (Eden et al, 1983) and has been increasingly used as a fully fledged organizational Group Decision Support System (Eden and Ackermann, 1992; Ackermann et al., 1993; Ackermann and Eden, 2001) and for strategy development and implementation within both public and private sectors (Eden and Ackermann, 1998). This particular approach of cognitive mapping is based on ‘‘personal construct theory’’ (Kelly, 1955) and has been developed, following extensions to the use of ‘‘Repertory Grids’’, for the purpose of capturing a ‘‘personal construct system’’. The cognitive mapping approach proposed by Damart (2008) can give information about the contribution of each participant to the exploration of a problem, in order to help the facilitator to form relevant participant sub-groups. A cognitive mapping procedure is developed in (Norese, 1995; Buffa et al, 1996) and as a tool of MACRAME, a problem structuring method. Two different applications, that are proposed in (Norese and Salassa, 2010 and Danna, 2011) were analyzed. In the third chapter methods of “actors analysis” are identified and analyzed in relation to problems context characterized by multiple actors, perspectives, experiences and competing interests. They can be proposed in order to study the structure of the decision context where the individual/organizational actors (or the potential actors) play a role (or multiple roles) and activate relationships, to analyze their points of view and identify new potential actors or to understand and reduce organizational complexity. In literature there are approaches that come from different disciplines with the aims to identify and analyze actors in a socio-technical system or in a process and to study the relations between them. Understanding who are the actors that have an “interest” and a possible role in decision making processes and can hinder essential information or enrich whole knowledge is important in a socio-technical system (De Bruijn, et al, 2002). Learning about the actors’ different problem perceptions and different content aspects of the problem situation is essential to estimate who are the “enemies” and who are “friends”, whom do you need or don't need. Actor analysis is also essential to legitimate a problem formulation in which different actors recognize themselves (Van der Poel, 1993). Approaches, methodologies and methods are proposed in different fields, employ different theoretic perspectives and focus on different aspects of multi-actor processes. Their use implies different expertises and in some cases a lot of time. All the findings and discussions of these analyses may be of relevance in all the domains where there are dimensions of multi-actor and organizational complexities. The integration of actors analysis methods and cognitive mapping techniques could produce knowledge elements that are useful to clarify the cognitive aspects and complete the vision of the whole situation. Actor analysis methods can be used to analyze and understand the context, but also to identify new actors or actors who could be involved, and to capture and represent differing perceptions of the situation and some problem formulations. The knowledge elements that come from the cognitive maps could also be used to improve the actor analysis (proposing, for exempla, new activities or actors, and related complex issues, to be analyzed). The actor network knowledge can facilitate the analysis of the actors’ needs, when these propose less clear or contradictory concepts that have to be analyzed by cognitive maps. The effectiveness of each method can be improved by the integration of a “complementary” method. When this integration is made and a clearer idea of the problem is acquired, some classical OR methods (in particular, in this research, mathematical programming and multiple criteria decision analysis) can be integrated, as in a second integration step, in order to structure the acquired knowledge elements in models. This second integration is made in order to transform concepts and relationships, that are structured and synthesized by actor networks and cognitive maps, into models (with actions, criteria and parameters, or objectives, variables, constraints and parameters) and to apply methods in order to elaborate possible solutions and compare them, or modify the models and identify new aspects and points of view, in decision contexts that include key actors. “Cycling between modeling approaches gave benefits that could not have been attained by either hard or soft modeling in isolation” (Ackermann et all., 1997). This “two steps” integration process is not linear but, as a learning cycle, better defines the knowledge elements that are essential in an innovation process and improves communication and modeling. The effectiveness of the interdisciplinary integration is tested on a real application, SMAT project. Procedure and results are shown in the fourth chapter. SMAT project was activated in January 2009 as the first phase of the global project of a new advanced system to monitor the territory (the meaning of the acronym SMAT) and it was financed by a public institution, the Piedmont Region. The project involved several enterprises and some research units of the Politecnico di Torino (with different skills, from different departments as DIASP, DELEN, DISPEA, DAUIN) and the University of Turin, under the leadership of Alenia Aeronautica, a Finmeccanica company, which is is active in the aeronautical military and civil markets. The purpose of SMAT-F1 is to test the performances of some legacy Unmanned Aerial Vehicles (UAVs), working as an integrated monitoring system, and to identify all the specific innovations that have to be introduced in the UAVs, that have to be equipped with sensors for specific data acquisition uses and communication systems for data transmission even in critical situations, in order to guarantee a civil use of this new monitoring system. The control stations of each UAV and the central control station have to ensure the coordination of the global system. The central control station has to guarantee a good interface between the system and the end users. In the first phase of the project, the aim of the research group of the DISPEA department was the identification of the organisations that could be the clients of this new monitoring service and the analysis of their monitoring needs, for the test of three legacy UAVs, but above all for the future phases of the SMAT project, when the new system has to be designed and implemented. In this case, some potential users of the SMAT technologies are the key actors of the land monitoring processes and recording and tracking their points of view could be important to understand the situation and also to involve some of them, in the future, in a decisional structure that should facilitate the design of both the innovative system and the new monitoring service. A correct acquisition of these points of view needs understanding and control of some uncertainties that are connected to the validity of the acquired knowledge (in terms of reliability, generability, global consistency, completeness,..), the value systems of the key actors, at individual and organizational level, and the nature of the relationships between the organizations that are involved in land monitoring processes. Open interviews starting from a framework of key questions can be more useful than a questionnaire to underline and clarify these uncertainty elements and to obtain an idea of what the knowledge elements that have to be acquired and analysed are. An integrated procedure was developed in order to define problem formulations, model frameworks and parameters starting from the structured, partially structured and unstructured knowledge elements that the survey has proposed. The procedure started from the validated interviews, and the suggestions from experts and literature in order to identify all the different monitoring needs. The texts of the validated interviews were analysed and all the structured knowledge elements (land monitoring current activities and characteristics such as cost and required quality; monitoring needs and factors that should characterize the new monitoring activities) were acquired and introduced in a data base and in some lists. Then a clustering approach was used for the definition of five land monitoring categories (or macro activities) and the assignment of all the expressed monitoring needs to these categories. A data base of all the possible uses of SMAT, with the elements that sprung from the interviews but also the points of view of the experts, was used to test the completeness of the set of macro activities and the consistency of the clustering approach, where the proximity of the needs, that were expressed during the interviews in relation to the five identified categories, was maximized and the number of the basic activities that synthesize similar needs was minimized. The last activity with the structured knowledge elements consisted in the definition of the main parameters that allow the basic activities to be described and treated in the models that the integrated analysis of the partially structured and unstructured knowledge elements is able to propose.