Review of literature between 1968-2012: Methods of Decision Making Concerning Risk Management

Review of Literature between 1968-2013: Methods of Decision Making Concerning Risk Management  

Introduction

The petrochemical industry on the southwest coast of Iran is the primary producer of petrochemical products in Iran. On 2 August 2012, a vapour cloud explosion happened in this company as a result of a hydrocarbon gas leakage from input pipelines. Those in charge of production and even the firefighters failed to notice the release. When a fire engine was then started, which is a routine task to ensure readiness in case of a need for firefighting, the hydrocarbon cloud exploded and spread to other parts of the site. Consequently, this caused significant numbers of injuries and deaths, uncontrolled hydrocarbon flow, air pollution and stopped some production units for one month. Although operating companies and regulators tightly control the activities in these kind of industries to prevent disasters [1], why do petrochemical industries and other oil and gas industries repeatedly experience this kind of accident?

Decision making is critically important to the safe and efficient running of complex socio-technical systems [2]. In any organization that operates complex technology, for instance nuclear power plants, oil, gas, and petrochemical industries, the interpretation of events highlights that accident scenarios can be varied but the type of barrier failures are common. One probable failure is wrong decision making which can happen in different circumstances and by different people at different levels of a company.

The ultimate aim of risk management and decision makers is to find the potential failures in particular circumstances and to take a different course of action to avoid accidents. In the context of decision making, we focused on the following questions: What are the common methods of decision making? Which of them are applied in risk management? What are the facilitators and obstacles of these methods? The implications of these findings, together with the recommendations, will be discussed below.

Material and methods

We reviewed and summarized the more relevant published studies to identify a number of different models and potential facilitators and obstacles for the implementation of decision making in risk management. Several databases were used for the literature search: Google scholar, Scopus and Web of Knowledge (Science). The terms ‘decision making’, ‘models’, ‘risk management’, ‘safety decision making’ in combination with ‘strategic’, and ‘operational decision making’ were used. The bibliographies of selected references were also searched. More than 220 articles written from 1968 to 2012 were found, of which 180 met the inclusion criteria. Studies were excluded if they described decision making in domains that are very far from safety-related decision making, such as students, patients or artificial decision making by computer. The full papers were reviewed and summarized by the first author. Papers were categorized according to the following criteria: I) decision making in general, containing individual and group decision-making models, influencing factors or biases of decision making and some decision support systems (n=140); II) safety decision making (n=36); and III) management and decision making (n=32), and within these groups, 11 papers indicated a link between safety, decision making, and management.

Results

The reviewed studies show that there is a considerable scientific literature about different models of decision making and the diverse factors affecting decision makers in other domains. However, there are few applications or assessments of these models in risk management. The results also indicate that decision making is a new scope in risk management as you can see in Table 1. Most of these studies have been done after 2000.

Table 1. Decision models in risk management according to methodology, aim of research and description.

*Abbreviations:

MCDM: Multi-criteria decision making AHP: Analytical hierarchy process

ANP: Analytic network process method NDM: Naturalistic decision making

RPD: Recognition primed decision making FTA : Fault tree analysis

TOPSIS: Technique for order preference by similarity to the ideal solution

DEMATEL: A decision-making trial and evaluation laboratory (DEMATEL)

In what follows, the application of different decision models in risk management in previous studies is discussed.

Safety decision problems are characterized by a number of alternatives, various criteria which are sometimes immeasurable, different decision makers with conflicting objectives and complex relations, as well as uncertain outcomes, so multi-criteria decision making is necessary in safety management. Multi-Criteria Decision Making (MCDM) is a branch of operation research models which concerns decision problems in the presence of multiple and conflicting criteria. This major class of models is divided into multi-objective decision making (MODM) and multi-attribute decision-making (MADM). MODM is designed to deal with two or more conflicting objective planning problems which can be solved by designing the best alternative, while MADM methods are more adequate for choosing the best option among predetermined finite alternatives [4]. Furthermore, the criteria for judgment of multi-objectives are objectives while the criteria to test acceptability in MADM are attributes.

There are several methodologies for multi-attribute decision making. The analytic hierarchy process (AHP), initially described by Saaty [5], is a structured multi-attribute decision method used in complex decision making and is the most widely used of the multi-criteria comparison methods in risk assessment [6] . The AHP method has been applied to analyse stochastic and fuzzy cases of uncertainties [7] and some probabilistic multi-criteria decision-making methods are related to AHP [8].

In AHP, a matrix of pair way comparison is formed to provide the relative preferences of the decision maker for various options. The logged pairwise comparison must be independent and normally distributed with a constant variance [8]. AHP uses both qualitative and quantitative data and is applicable for decision situations involving subjective expert judgments[9]. It decreases the inconsistency of expert judgments. It is a reliable method for comparing risk factors, evaluating risks, defining priorities, designating resources and measuring performance [9]. The AHP models human factors more clearly, simply and practically .

The fuzzy MCDM model is another widespread method of decision making which has been developed to aggregate the decision maker’s subjective assessment about the appropriateness of alternatives, criteria selection and criteria weighting. Fuzzy decision models are appropriate where the decision maker is unable to express his/her preferences due to lack of sufficient knowledge of the problem or inability to distinguish the degree to which one alternative is preferred over others. Fuzzy approaches enable the group to identify the similarities and differences of their judgments. They incorporate imprecise, vague and uncertain data which are characteristics of human-centred activities into an analysis [10]. The fuzzy multi-criteria group decision-making approach is suitable when a large number of attributes are considered in the evaluation [11]. Another important characteristic of the model is its ability to manage the complexity of multi-dimensional entities properly and to integrate different kinds of information [10].

The above-mentioned models are the most important methods of rational decision-making models applied in risk management; however, Klein has argued that there are some limitations to the analytical decision strategies[12]. First, if they are used in improper conditions where rapid and efficient reactions are crucial, rational models might leave the decision maker incapable of mitigating outcomes. Second, in risk management, managers should be actively involved in decisions to shape events and not passively awaiting the outcomes [13]; consequently, naturalistic decision-making (NDM) research emerged in the 1980s to study how people actually make decisions in real-world circumstances. Lipshitz identified nine NDM models that had been advanced in parallel[14]. These models and theories were developed from rejection of subjective expected utility (SEU) theory and decision research which was based on laboratory settings [15].

NDM models focus on the role of the decision-maker’s experience to rapidly recognize and categorize the situations and to make a decision where other plausible alternatives exist even though the decision maker does not notice or compare them. NDM approaches are knowledge-based approaches and they elaborate the decision-making process into a prior stage of perception and recognition of situations [13]. These approaches emphasize the hazardous and complex environment in which time pressure or other influencing factors mitigate the suitability of rational decision making[16]. In total, naturalistic decision making contrasts with rational decision making in the process of identifying a set of alternatives, evaluating those alternatives, weighting the criteria, rating the options, and selecting the final option with the highest score.

Among NDM models the Recognition-primed decision making (RPD) model lies within the framework of cognitive task analysis [16]; it focuses on mental simulation and situation diagnosis. Pattern recognition by an expert decision maker is one fundamental process in this model. It means that the decision maker compares a situation with past experiences and recaptures a potential course of action which has been successful in the past [17]. The RPD model attempts to demonstrate how the decision maker actually decides under time pressure, with ill-defined goals, vague information and changing conditions [18].

RPD differs from classical decision making by a number of characteristics. First, it focuses on a situation assessment instead of option judgment to contrast the strengths and weaknesses of different alternatives. Second, it relies on recognition of a first good option by an experienced decision maker rather than the generation of many alternatives. It is based on satisfaction rather than optimization[19]. Finally, RPD enables the decision maker to continually prepare to take an action based on the current evaluation; however, in rational decision making, the decision maker should wait until finding out which option is rated the highest [18]

Conclusion

It is important to stress the scope of the work in this paper. We have provided an initial analysis of decision-making models and methods that have been conducted in risk management for launching a successful approach to safety-critical decision making in an organization. The results of the review determine important factors influencing decision making such as task goals, contexts, and decision processes. We also have to acknowledge that there is no single effective decision-making strategy or procedure for balancing different concerns. We should accept the limitations of the tools, and use multiple methods for improvement of the decision making.

Among different models of rational decision making, Analytical Hierarchy Process is the most popular technique in safety management, followed by fuzzy sets to tackle uncertainties in the data and NDM models to find what people actually decide in real conditions. Application of multiple methods particularly AHP and fuzzy sets, and enhancement of interactive decision support systems is observed in the literature; nonetheless a validation of results especially with multiple methods has not yet been performed.

In AHP subjective preference, objective information and expert knowledge can be involved in the evaluation of alternatives, so it is a useful tool to tackle complexity and ill-defined goals; however, AHP does not provide an in-depth analysis of options particularly in intrinsic uncertainty of data. Another limitation is when the number of criteria or alternatives increases, the number of pair way comparisons expands rapidly [7]. Finally, data entry errors, missing information, poor construction, and modelling problems can cause inconsistency in this model. A new trend in the application of fuzzy sets has been observed in risk assessment to improve the precision of risk assessment. Fuzzy logics provides a framework for modelling imprecise and ambiguous criteria. This approach is not sensitive to small changes, and a dynamic transition from one category to the others is possible.

Comparison of analytical decision making with descriptive decision making shows that both methods have special obstacles and facilities. For instance, if analytical decision methods are utilized in wrong conditions, decision makers are unable to react effectively and quickly. In contrast, if descriptive methods like RPM are misapplied, they can leave the decision maker incapable of identifying the course of actions which is needed[20]. These conditions can occur when the decision maker lacks the ability to mentally simulate alternatives to find pitfalls or to fail to optimize, so a recognitional approach is useful when the decision maker is experienced, time pressure is greater, or conditions are less stable. Conversely, analytical decisions predominate when the available data are abstract, the problems are complicated and mixed, a dispute between different agencies exists or justification of the course of action is necessary.

Acknowledgement

A preliminary version of this paper was presented at the Congress ORP conference in 2014.

 References

[1] J. Hayes, “Operator competence and capacity – Lessons from the Montara blowout,” Saf. Sci., vol. 50, no. 3, pp. 563–574, Mar. 2012.

[2] D. P. Jenkins, N. a. Stanton, P. M. Salmon, G. H. Walker, and L. Rafferty, “Using the Decision-Ladder to Add a Formative Element to Naturalistic Decision-Making Research,” Int. J. Hum. Comput. Interact., vol. 26, no. 2–3, pp. 132–146, Mar. 2010.

[3] E. Topuz, I. Talinli, and E. Aydin, “Integration of environmental and human health risk assessment for industries using hazardous materials: a quantitative multi criteria approach for environmental decision makers.,” Environ. Int., vol. 37, no. 2, pp. 393–403, Feb. 2011.

[4] G. a. Mendoza and H. Martins, “Multi-criteria decision analysis in natural resource management: A critical review of methods and new modelling paradigms,” For. Ecol. Manage., vol. 230, no. 1–3, pp. 1–22, Jul. 2006.

[5] T. L. Saaty, “How to make a decision : The Analytic Hierarchy Process,” Eur. J. Oper. Res., vol. 48, pp. 9–26, 1990.

[6] A. Badri, S. Nadeau, and A. Gbodossou, “Proposal of a risk-factor-based analytical approach for integrating occupational health and safety into project risk evaluation,” Accid. Anal. Prev., vol. 48, pp. 223–234, 2012.

[7] J. Kangas and A. Kangas, “Multiple criteria decision support in forest management—the approach, methods applied, and experiences gained,” For. Ecol. Manage., vol. 207, no. 1–2, pp. 133–143, Mar. 2005.

[8] E. D. Hahn, “Link function selection in stochastic multicriteria decision making models,” Eur. J. Oper. Res., vol. 172, no. 1, pp. 86–100, Jul. 2006.

[9] R. D. ; Henderson and S. P. Dutta, “Use of the Analytic Hierarchy Process in ergonomic analysis,” vol. 9, pp. 275–282, 1992.

[10] M. Omero, L. D’Ambrosio, R. Pesenti, and W. Ukovich, “Multiple-attribute decision support system based on fuzzy logic for performance assessment,” Eur. J. Oper. Res., vol. 160, no. 3, pp. 710–725, Feb. 2005.

[11] H. Wang, “Fuzzy multicriteria decision making – an overview,” J. Intell. Fuzzy Syst., vol. 9, pp. 61–83, 2000.

[12] G. Klein and D. Klinger, “Naturalistic Decision Making,” Hum. Syst. IAC Gatew., vol. 4, no. 1, pp. 1–4, 1991.

[13] G. Klein, “Naturalistic Decision Making,” Hum. Factors J. Hum. Factors Ergon. Soc., vol. 50, no. 3, pp. 456–460, Jun. 2008.

[14] R. Lipshitz, G. Klein, J. Orasanu, and E. Salas, “Taking stock of naturalistic decision making,” J. Behav. Decis. Mak., vol. 14, no. 5, pp. 331–352, Dec. 2001.

[15] J. Gore, “Naturalistic Decision Making and Organizations: Reviewing Pragmatic Science,” Organ. Stud., vol. 27, no. 7, pp. 925–942, Jul. 2006.

[16] G. Lintern, “A Comparison of the Decision Ladder and the Recognition-Primed Decision Model,” Jourmal Cogn. Eng. Decis. Mak., vol. 4, no. 4, pp. 304–327, 2010.

[17] E. Salas, M. a. Rosen, and D. DiazGranados, “Expertise-Based Intuition and Decision Making in Organizations,” J. Manage., vol. 36, no. 4, pp. 941–973, Oct. 2009.

[18] K. E. T. Al, “Decision making in action : models and methods,” vol. 8, pp. 250–253, 1996.

[19] H. A. Simon and B. H. A. Simon, “Rational Decision Making in Business Organizations,” vol. 69, no. 4, pp. 493–513, 1979.

[20] G. A. Klein and R. Calderwood, “critical Decision Method for Elicitinting Knowledge,” IEEE Trans. Syst. Man Cybern., vol. 19, no. 3, 1989.

[21] J. J. H. Liou, G.-H. Tzeng, and H.-C. Chang, “Airline safety measurement using a hybrid model,” J. Air Transp. Manag., vol. 13, no. 4, pp. 243–249, Jul. 2007.

[22] I. Linkov, F. K. Satterstrom, G. Kiker, C. Batchelor, T. Bridges, and E. Ferguson, “From comparative risk assessment to multi-criteria decision analysis and adaptive management: recent developments and applications.,” Environ. Int., vol. 32, no. 8, pp. 1072–93, Dec. 2006.

[23] J. K. Levy, “Multiple criteria decision making and decision support systems for flood risk management,” Stoch. Environ. Res. Risk Assess., vol. 19, no. 6, pp. 438–447, Oct. 2005.

[24] T. Zhijun, Z. Jiquan, and L. Xingpeng, “GIS-based risk assessment of grassland fire disaster in western Jilin province, China,” Stoch. Environ. Res. Risk Assess., vol. 23, no. 4, pp. 463–471, Apr. 2008.

[25] S. Saffarian, M. Shafiee, and N. Zaredar, “A Novel Approach Toward Natural and Anthropogenic Risk Assessment of Gas Power Plants,” Hum. Ecol. Risk Assess. An Int. J., vol. 20, no. 2, pp. 346–365, Mar. 2014.

[26] A. Hatami-Marbini, M. Tavana, M. Moradi, and F. Kangi, “A fuzzy group Electre method for safety and health assessment in hazardous waste recycling facilities,” Saf. Sci., vol. 51, no. 1, pp. 414–426, Jan. 2013.

[27] Y. Peng, Y. Zhang, Y. Tang, and S. Li, “An incident information management framework based on data integration, data mining, and multi-criteria decision making,” Decis. Support Syst., vol. 51, no. 2, pp. 316–327, May 2011.

[28] U. Topacan, A. N. Basoglu, and T. U. Daim, “AHP application on evaluation of health information service attributes,” Int. Conf. Manag. Eng. Technol., pp. 486–493, Aug. 2009.

[29] M. V. Frank, “Choosing among safety improvement strategies: a discussion with example of risk assessment and multi-criteria decision approaches for NASA,” Reliab. Eng. Syst. Saf., vol. 49, no. 3, pp. 311–324, Jan. 1995.

[30] H. S. Sii, T. Ruxton, and J. Wang, “A fuzzy-logic-based approach to qualitative safety modelling for marine systems,” Reliab. Eng. Syst. Saf., vol. 73, no. 1, pp. 19–34, Jul. 2001.

[31] C. W. Johnson, B. Kirwan, A. Licu, and P. Stastny, “Recognition primed decision making and the organisational response to accidents: Überlingen and the challenges of safety improvement in European air traffic management,” Saf. Sci., vol. 47, no. 6, pp. 853–872, 2009.

[32] F. Xhafa, E. Asimakopoulou, N. Bessis, L. Barolli, and M. Takizawa, “An Event-based Approach to Supporting Team Coordination and Decision Making in Disaster Management Scenarios,” 2011 Third Int. Conf. Intell. Netw. Collab. Syst., pp. 741–745, Nov. 2011.

[33] M. C. R. V. Paulo V.R. Carvalho,, Isaac L. dos Santos, “Nuclear power plant shift supervisor’s decision making during microincidents,” Int. J. Ind. Ergon., vol. 35, no. 7, pp. 619–644, Jul. 2005.

[34] a. J. Brito and A. T. de Almeida, “Multi-attribute risk assessment for risk ranking of natural gas pipelines,” Reliab. Eng. Syst. Saf., vol. 94, no. 2, pp. 187–198, Feb. 2009.

[35] G. Reniers and Y. Pavlova, “Game-theory for safety and security in the chemical industry.” Springer, 2010.