Efficiency-guided method for traveling-salesman-like resource allocation problems
William H. Press, Curtis G. Callan · Journal of Guidance Control and Dynamics · 1988
In resource allocation problems, it is possible to define the local efficiency of a small change, namely the ratio of change in value to marginal cost. This efficiency can be used to guide the exploration of an otherwise stochastic method like simulated annealing. An algorithm embodying this idea, termed efficiency-guided addition, subtraction, and permutation, (EGASP), is described for problems akin to the traveling salesman problem.