An Incremental Algorithm for Minimal Joint Graph Synthesis

Abstract

The paper provides the description of an incremental algorithm that accelerates a minimal joint graph upgrade when a new node with a correct load is inserted. The algorithms correctness is proven. We compare relative performance statistical estimates for the new incremental algorithm and two known greedy and streightforward algorithms that implement minimal joint graph synthesis. The computational experiments results are represented with plots, discussed and summarized. For the middle and large sized sets of loads, the incremental algorithm is 3-6 times faster than the streight-for-ward one and 10-50 times faster than the greedy one. For the small sized sets of loads, the algorithms permormance differs less dramatically. The relative performance variation of compared algorithms are the larger the higher fraction of loads with 5-9 symbols the set of loads has. The paper also reaches didactical goals: in the context of minimal joint grah synthesis, we demonstrate a rational application of the algorithm incrementation, an example of relative algorithmical performance statistical estimates, an approach to the visualisation of performance comparison compuational experiments.