An SMAA-based approach for multi-attribute fixed resource allocation problem with random attribute values and uncertain weights

Purpose In a typical problem of allocating a fixed resource across a set of decision-making units (DMUs), the decision maker will decide how much resource to allocate to each DMU. Real-world decision-making contexts frequently pose challenges in precisely assessing each DMU’s performance on distinct parameters and in ascertaining the relative importance of these parameters, since resource allocation often precedes the decision-making process. Conventional decision-making approaches fall short in effectively addressing such complexities. In this paper, the authors propose a stochastic multi-attribute acceptability analysis (SMAA)-based model that can deal with the resource allocation problem with uncertain attribute values and unknown attribute weights. Design/methodology/approach The SMAA-based model first regards the resource to be allocated as an evaluation attribute and then proposes a fair resource allocation solution that minimizes the gap between the holistic acceptability of each DMU. To address the optimization of the SMAA-based model, two heuristic strategies were devised: a modified particle swarm optimization (PSO) approach and an adaptive algorithm, both designed to yield acceptable outcomes within a constrained time frame. Findings The resource allocation model based on SMAA can provide an effective analytical framework for the resource allocation problem of different DMUs in uncertain environments. Meanwhile, the PSO algorithm and adjustment algorithm proposed in this paper can help decision-makers quickly find a fair resource allocation plan. Originality/value Primarily, this study introduces a novel approach rooted in SMAA for tackling resource allocation dilemmas where attribute values are uncertain and weight information is absent. Secondly, a fairness-centric strategy that minimizes the disparity in overall acceptability among various DMUs is presented. Thirdly, two heuristic algorithms, namely, PSO algorithm and adjustment algorithm, are developed to solve the problem.