The effectiveness of reason-based approaches for distribution network desig

Many operations management (OM) decisions are made in practice with only limited use of operations research tools. Instead, non-optimizing methods which rely on human reasoning and problem structuring abilities are often employed. This research assesses the effectiveness of such a method for distribution network design decisions.;Distribution network design is of key importance to a firm's strategic posture and potential for success. While there have been considerable advances in hardware and optimization software technology for network design decisions, existing optimization algorithms are not widely used and have not kept pace with the complexity of problems facing managers. Consequently, intuition, scenario evaluation, and heuristics are widely used.;The research reported here had three objectives: (1) to assess the relative effectiveness of humans using reason based approaches for facility location decisions of varying complexity, (2) to identify relationships between problem characteristics, decision maker attributes and problem solving performance, and (3) to identify effective search strategies for the scenario evaluation process. An experimental analysis of industrial managers' problem solving performances designed to achieve these objectives is described. Experimental factors included problem structure, problem size, and problem solving method. Associations between performance and subjects' demographic and cognitive attributes were also explored. Analytic techniques included ANOVA, factor analysis and multiple regression.;Performance measures were assessed relative to optimal solutions. The results indicate that the type of network problem solved has little effect on problem solving performance. Instead, the absolute number of decision variables contained in the problem is a better predictor of performance. The strength of this effect diminishes with problem size.;The pattern of solution search does not significantly affect ultimate solution quality. Decision maker experiences and usage of data are significantly associated with performance. However, all of these effects are mitigated by the availability of incremental change cost information, which provides guidance in the search process. The availability of this information dominates all other problem or user related effects.;These findings suggest that decision makers are capable of generating high quality solutions quickly using reason based approaches. However, the availability of even a small amount of guidance decreases the variability in problem solving performance significantly. Consequently, researchers should focus on extending the computational efficiency of existing algorithms and developing user friendly heuristics and optimizing techniques.