Adaptation on rugged landscapes
Management Science
Modularity and Innovation in Complex Systems
Management Science
Population Learning in a Model with Random Payoff Landscapes and Endogenous Networks
Computational Economics
Global optima results for the Kauffman NK model
Mathematical Programming: Series A and B
Phase transition in a random NK landscape model
Artificial Intelligence
| Hi-index | 0.00 |
Problems are often too complex to solve them in an optimal way. The complexity arises from connections between their elements, such that a change in one element influences the performance of other elements. Kauffman's NK model offers a way to depict such interdependencies and has therefore often been used in economic investigations of the influence of problem or search decomposition on the attainable results. However, papers on the effect of different decompositions on solution quality come to contradictory conclusions. Some observe an initial advantage of over-modularization where others do not. As they also differ in the employed search procedures, but do not base them on empirical findings, the present paper examines the results of more empirically based search strategies. Using algorithms based on innovation strategies derived from patent data, the paper establishes a clear advantage of correct problem decompositions.