Development of Algorithms for Decision Analysis with Interval Information

Optimization-assisted personalized event recommendation for event-based social networks

Since the emergence of Artificial Intelligence as a field, planning under uncertainty has been viewed as one of its crucial subareas. Accordingly, the current lack of scalability of probabilistic planning techniques is a major reason why the grand vision of AI has not been fulfilled yet. Besides being an obstacle to advances in AI, scalability issues also hamper the applicability of probabilistic planning to real-world problems. A powerful framework for describing probabilistic planning problems is Markov Decision Processes (MDPs). Informally, an MDP specifies the objectives the agent is trying to achieve, the actions the agent can perform, and the states in which it can end up while working towards the objective. Solving an MDP means finding a policy, i.e. an assignment of actions to states, that allows the agent to achieve the objective. Optimal solution methods, those that look for the “best” policy according to some criterion, typically try to analyze all possible states or a large fraction of them. Since state space sizes of realistic scenarios can be astronomical, these algorithms quickly run out of memory. Fortunately, the mathematical structure of some classes of MDPs has allowed for inventing more efficient algorithms. This is the case for Stochastic Shortest Path (SSP) problems, whose mathematical properties gave rise to a family of algorithms called Find-and-Revise. When used in combination with a heuristic, the members of this family can find a near-optimal, or even optimal, policy while avoiding the analysis of many states. Nonetheless, the sheer number of states in real-world problems forces algorithms based on state-level analysis to exhaust their capabilities much too early, calling for a fundamentally different approach. Moreover, several expressive and potentially useful MDP classes are currently not known to have a mathematical structure for creating efficient approximation techniques. This dissertation advances the state of the art in probabilistic planning in three complementary ways. For SSP MDPs, it derives a novel class of approximate algorithms based on generalizing state analysis information across many states. This information is stored in the form of basis functions that are generated automatically and the number of which is much smaller than the number of states. As a result, the proposed algorithms have a very compact memory footprint and arrive at high-quality solutions faster than their state-based counterparts. In a parallel effort, the dissertation also builds up mathematical apparatus for classes of MDPs that previously had no applicable approximation algorithms. The developed theory enables the extension of the powerful Find-and-Revise paradigm to these MDP classes as well, providing the first memory-efficient algorithms for solving them. Last but not least, the dissertation will apply the proposed theoretical techniques to a large-scale real-world problem, urban traffic routing being one of the candidates.

Human decisions about the design, deployment, and governance of AI systems are increasingly complex, involving multiple desiderata such as accuracy, fairness, transparency, cost, and environmental impact. This paper proposes a novel and scalable framework for managing this complexity through application of resources from the field of Multi-Criteria Decision Analysis (MCDA). After an introduction to the “ n > 2 Criteria Problem” in decision-making about AI, an introduction to MCDA methods, and a review of prior applications of MCDA to decision support in decisions involving AI, we lay out our own novel and scalable framework for applying MCDA to such decisions. We illustrate the framework through application to four simulated scenarios involving a medical institution’s choice among AI-powered diagnostic tools with varying performance on accuracy, fairness, transparency, and cost. We apply five multi-criteria decision analysis (MCDA) methods — namely, the Weighted Sum Model (WSM), the Analytic Hierarchy Process (AHP), the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), the VIKOR method (VlseKriterijumska Optimizacija I Kompromisno Resenje), and the Preference Ranking Organization METHod for Enrichment of Evaluations (PROMETHEE) — to rank ten AI models under four distinct weight configurations reflecting realistic deployment contexts. Sensitivity analyses are then conducted to examine how variations in criterion weight assignments affect model rankings across methods and scenarios. In concluding sections we discuss the scalability of these techniques for design, deployment, and policy decisions involving AI, as well as the likely limitations of MCDA methods for decision support involving AI.

Rough set theory has been proposed by Pawlak in the early 80s to deal with inconsistency problems following from information granulation. It operates on an information table composed of a set U of objects described by a set Q of condition and decision attributes. Decision attributes make a partition of U into decision classes. Basic concepts of rough set theory are: indiscernibility relation on U, lower and upper (rough) approximations of decision classes, dependence and reduction of attributes from Q, and decision rules induced from rough approximations of decision classes. The original rough set idea was failing, however, to handle preferential ordering of domains of attributes (scales of criteria), as well as preferential ordering of decision classes. In order to deal with multiple criteria decision problems a number of methodological changes to the original rough set theory were necessary. The main change is the substitution of the indiscernibility relation by a dominance relation, which permits approximation of ordered sets. In multiple criteria decision context, the information table is composed of decision examples given by a decision maker. The Dominance-based Rough Set Approach (DRSA) applied to this information table results with a set of decision rules, being a preference model of the decision maker. It is more general than the classical multiple attribute utility model or outranking model, and it is more understandable because of its natural syntax. In this chapter, after recalling the classical rough set approach and DRSA, we review their fuzzy set extensions. Moreover, we characterize the dominance-based rough approximation of a fuzzy set, and we show that the classical rough approximation of a crisp set is its particular case. In this sense, DRSA is also relevant in the case where preferences are not considered, but just a kind of monotonicity relating values of different attributes is meaningful for the analysis of data at hand. In general terms, monotonicity concerns relationship between different aspects of a phenomenon described by data: for example, “the larger the house, the higher its price” or “the closer the house to the city centre, the higher its price”. In this perspective, DRSA gives a very general framework for reasoning about data using only monotonicity relationships.

Multi-criteria screening of carbon dioxide utilization products combined with process optimization and evaluation

Rough sets theory for multicriteria decision analysis

Disposal of healthcare waste is a key issue of environmental sustainability in the world. The amount of healthcare waste is increasing every day, and it is necessary to adequately dispose of this kind of waste. There are various treatments for healthcare waste disposal, of which incineration of healthcare waste is one of the solutions. This paper suggests a model for selection of the type of incinerators that best solve the problem of healthcare waste in secondary healthcare institutions in Bosnia and Herzegovina. In the selection of incinerators, extended sustainability criteria were applied. Basic sustainability criteria: environmental, economic, and social criteria, were extended with the technical criterion. To assess which of the incinerators best meets the needs for healthcare waste collection, multi-criteria decision-making was used. For this purpose, a combination of two MCDA methods was applied in this paper, namely full consistency method (FUCOM) and compromise ranking of alternatives from distance to ideal solution (CRADIS). The FUCOM method was applied to determine the weights of the criteria, while the CRADIS method was applied to rank the alternatives. The best alternative of the six alternatives used is A2 (I8-M50), followed by alternative A1 (I8-M40), while the worst ranked alternative is A5 (I8-M100). These results were confirmed by applying the other six methods of multi-criteria analysis and the performed sensitivity analysis. The contribution of this paper is reflected through a new method of multi-criteria analysis that was used to solve decision-making problems. This method has shown simplicity and flexibility in operation and can be used in all problems when it is necessary to make a multi-criteria selection of alternatives.

Multi-criteria analysis

Uncertainty propagation and risk analysis oriented stochastic multi-criteria decision making for unconventional water resources management

In this paper, a hybrid model consisting of the fuzzy ARTMAP (FAM) neural network and the classification and regression tree (CART) is formulated. FAM is useful for tackling the stability---plasticity dilemma pertaining to data-based learning systems, while CART is useful for depicting its learned knowledge explicitly in a tree structure. By combining the benefits of both models, FAM---CART is capable of learning data samples stably and, at the same time, explaining its predictions with a set of decision rules. In other words, FAM---CART possesses two important properties of an intelligent system, i.e., learning in a stable manner (by overcoming the stability---plasticity dilemma) and extracting useful explanatory rules (by overcoming the opaqueness issue). To evaluate the usefulness of FAM---CART, six benchmark medical data sets from the UCI repository of machine learning and a real-world medical data classification problem are used for evaluation. For performance comparison, a number of performance metrics which include accuracy, specificity, sensitivity, and the area under the receiver operation characteristic curve are computed. The results are quantified with statistical indicators and compared with those reported in the literature. The outcomes positively indicate that FAM---CART is effective for undertaking data classification tasks. In addition to producing good results, it provides justifications of the predictions in the form of a decision tree so that domain users can easily understand the predictions, therefore making it a useful decision support tool.

Applying Multi-Criteria Decision Analysis to the Life-Cycle Assessment of vehicles

Selection of discrete multiple criteria decision making methods in the presence of risk and uncertainty

Recent studies have identified that multi-criteria group analysis methods should take the concepts of risk and uncertainty into account. In some real-life situations, determining the exact values for the potential alternatives' performances and criteria weights is so difficult. To overcome with these situations, their values should be regarded as fuzzy and fuzzy intervals. In this respect, interval-valued hesitant fuzzy sets (IVHFSs) as a suitable modern fuzzy sets theory can be considered because this theory allows decision makers (DMs) to assign some interval-values membership degrees for an alternative in terms of selected criteria under a set to margin of errors. Hence, this paper proposes a novel soft computing approach, namely IVHF-MCGA, based on new interval-valued hesitant fuzzy complex proportional assessment (IVHF-COPRAS) method that can be applied in solving the multi-criteria group decision-making (MCGDM) problems under uncertainty. In this approach, preference values of potential alternatives versus the selected criteria and weights of each criterion are expressed by linguistic variables and then are transformed to interval-valued hesitant fuzzy elements (IVHFEs). In addition, an interval-valued hesitant fuzzy entropy (IVHF-entropy) method is extended to determine the criteria weights by considering the DMs' opinions about the relative importance. Also, a new interval-valued hesitant fuzzy compromise solution (IVHF-CS) method is introduced to estimate the weight of each DM in the group decision-making process along with the last aggregation for the DMs' judgments to avoid the data loss. Then, three practical applications about the robot selection, industrial site selection and rapid prototyping process selection problems are considered to explain steps of the proposed IVHF-MCGA approach and to indicate its validity and applicability. Finally, a comparative analysis between the proposed approach and fuzzy group TOPSIS method is presented based on four comparison parameters, including adequacy to changes of alternatives and criteria, agility in decision process, influence of DMs' weights and impact of first and last aggregations, to show its suitability.

Machine ranking plays a crucial role in manufacturing, influencing both economic efficiency and technical performance. However, this task is inherently complex because each machine must be evaluated simultaneously across multiple, often conflicting, criteria. This challenge highlights the need for Multi-Criteria Decision-Making (MCDM) methods. This study proposes a new criteria weighting approach, the SITDER method, that aims to improve the stability of machine ranking results across different MCDM techniques. SITDER integrates two methods: Skewness Impact Through Distributional Evaluation (SITDE) and Rank Order Centroid (ROC). In the first stage, SITDE is used to determine the relative importance of the criteria. In the second stage, ROC is applied to compute the final criterion weights. The effectiveness of SITDER was validated using two case studies involving the ranking of 13 handheld polishers and 12 grinding machines. In both examples, the rank correlation coefficients among the MCDM methods were consistently higher when using SITDER compared to SITDE alone. In Example 1, the average Spearman coefficient increased from 0.6282 (SITDE) to 0.739 (SITDER). In Example 2, it increased from 0.9643 to 0.9744. These results demonstrate that SITDER is a highly effective weighting method and a valuable addition to the set of available criteria weighting techniques.

Multi-criteria decision analysis to select metrics for design and monitoring of sustainable ecosystem restorations