|
||
Planners
as Nonlinear and Complex Explorers
|
||
|
Planners as Odd Matchmakers Recombination One way that biological organisms explore their adaptive "space" is through sexual reproduction where recombinations of parental genetic material afford the opportunity for modifications that may prove more fit for the species. The computer scientist and complexity pioneer John Holland (1992) has created adaptive computer programs called genetic algorithms based on sexual reproduction as a paradigm of "crossing-over" or the mixing of genetic material (which become bit strings in his programs). The programs evolve by both sexual-like recombination as well as through random mutations; each new modification that is closer to the solution is given a heavier weight. Then the program over many generations converges to a solution. The computer scientists Gerhardt Bruderer and Martin Maiers along with the complexity management consultant and theorist Glenda Eoyang (Maiers and Eoyang, 1997) have been designing a genetic algorithm as a decision support tool for managers. This program can easily be modified for decision-making in planning as well. But again, such a usage is dependent on planners revising their view of what their main roles are to be. Recombination also comes up in Kauffman s N/K model. For Kauffman, sexual mating or reproduction allows a kind of "God's Eye" peek at the peaks (Kauffman, 1995). The genetic recombinations that result from sex between organisms at different locations on a landscape allows the adapting "population" to "look at" the regions between the parental genotypes. In this way recombination allows the adapting population to make use of large scale features of the landscapes to find high peaks. In fact, Kauffman found in his N/K landscapes that populations using mutation and recombination as well as selection improve far more rapidly than those using only mutation and selection (Kauffman, 1995). If the fitness landscape looks like the Alps, then the peaks carry mutual information about where to find high peaks: they are nearby! Moreover, if parents are high up on the peaks, then the kids will have a greater chance to start out higher. Yet, recombination can actually be harmful on a totally random landscape since if parents are at local peaks, recombination can lead progeny to be "dropped off" in a place with lower fitness. Furthermore, when a search is merely random with no clues about upward trends, the only way to find the highest pinnacle is to search the whole space. Recombination is going on in organizations in an unprecedented manner with the accelerating pace of mergers and acquisitions. Previously competitive organizations are now joined and the frequent issue concerns how these previously separate, even hostile entities can possible work together. Whereas the traditional approach might be to impose a new structure or plan or working procedures on the newly merged system, an approach informed by genetic algorithms or the N/K model would see this recombination and the potential conflict it might engender as a great opportunity for the emergence of new organizational practices and directions (see Goldstein, "Leadership and Emergence" in the File Cabinet section). Then the intervention would not be to dampen differences of opinion but to highlight them, amplify them and allow a more adaptive organizational structure to emerge as a result of the merger.
|
|
Next | Previous
Copyright © 2001, Plexus Institute. Permission |