Abstract
Decision support systems (DSS) are information systems that facilitate human decision-making through the presentation and analysis of data. Primarily, these information systems allow people to improve decisions by using additional data and information that they have access to but do not always use to make a more informed decision. DSS efforts enable military and civilian leaders to improve strategic plans and make decisions. Increasingly, DSS are seen in use by both the public and decision makers to make sense of big data, as seen with COVID-19 presentations. Artificial intelligence (AI) can be used to provide rapid interpretation of the raw data and results for use within DSS systems. DSS is colloquially termed a “dashboard” and involves three main components: the database, model, and the user interface. This article explores archetypes of DSS and aims to discuss each component in equal measure since ignoring one aspect leads to various issues (e.g., a DSS employing a good model with appropriate data but poor implementation).
TopIntroduction
Decision support systems (DSSs) are a sub-set of information systems that support human decision-making through computerized systems that provide contextual information. DSSs allow decision-makers to improve their strategic planning and management control. (Yong & Taib, 2009). With the continual expansion of big data (Bihl, Young II, & Weckman, 2016), the expansion of DSSs at the consumer to enterprise level and the increasing demands of senior leaders for an “end-to-end set of information capabilities” (Shelton, 2000), DSS’ need and use will continue to expand and grow. This will become an integral part of the data and information environment involving storing the data and information in a database/structure, understanding the information within a given model and presenting it to the user (user interface).
DSSs can be applied in a variety of areas to assist decision-makers as in controlling inventory, assessing consumer behavior, scheduling, forecasting, safety, planning, and risk assessment (Turban, Aronson, & Liang, 2008). From a business standpoint these systems have a wide-range of applications including political analysis (Berg & Rietz, 2003), investigating social implications (Turoff, Hiltz, Cho, Li, & Wang, 2002), medical/clinical decision making (Shaffer & Coustasse, 2012), developing educational programs (Tatnall, 2007), understanding consumer behaviors (Koufaris, Kambil, & Labarbera, 2001), cyber defense (Gutierrez et al., 2018), modeling and simulation result interpretation (Bihl et al., 2009, 2020), evaluating military decisions (Klimack, 2002), assessing environmental policies (Poch et al., 2004), forecast demand (Efendigil, Onut, & Kahraman, 2009), predict stock performance (Kuo, Chen, & Hwang, 2001), and understand power system loads (Santana, et al., 2012). DSSs can further be used as part of a simulation environment, Figure 1, wherein the DSS provides insights into the simulation of a larger system (Bihl et al., 2009). More recently, with the rise of artificial intelligence (AI) and concerns of the black-box nature of AI, DSS interfaces have become of interest to explain AI decisions and inferences (van der Waa et al., 2021; Wanner et al., 2020).
Figure 1. Example of a DSS for a satellite power simulation system, the SPECTTRA Viewer
DSS usage, design and development have expanded to near ubiquity with the emergence of business analytics; includes internet listservs, web directories, and Google searches (Lankton, Speier, & Wilson, 2012). This has expanded in the 2010s to include dashboard applications which are embedded in webpages (O'Brien & Stone, 2020). This expansion and use of DSSs became evident in the public sphere in the COVID-19 pandemic whereby a multitude of DSS systems became available online (Lan et al., 2021), an example of which is presented in Figure 2.
Figure 2. The State of Ohio’s COVID-19 Vaccine Dashboard
Key Terms in this Chapter
Active DSS: One explicitly recommending solutions to a decision maker.
Knowledge-Driven DSS: One designed to leverage expert knowledge on particular decision-making categories; decisions are supported by facts, rules, and procedures based on similar problem structures.
Communication-Driven DSS: One aiding a decision maker with decisions or tasks when there are many individuals working together.
Document-Driven DSS: One designed to manage, retrieve, and summarize information from various electronic file formats.
Passive DSS: One that does not explicitly recommend solutions, but aids a decision maker in the decision making process.
Structured Problems: A problem type with very well-known system relationships. In general, decisions made for structured problems are well defined and documented, requiring no expert knowledge.
Unstructured Problems: A problem type where system relationships are not known; in general, decisions made for unstructured problems are not defined and require expert knowledge.
Data-Driven DSS: One utilizing data with access being either single or a variety of databases.
Cooperative DSS: One designed in an iterative manner where recommended solutions are refined by decision makers and sent back for validation. The process is repeated until arriving at a consolidated solution.
Model-Driven DSS: A DSS utilizing mathematical models built from probability, statistics, and machine learning algorithms and strategies. These DSS manipulate data to assist decision makers.