Modell for flerscenario strukturoptimering - modellbeskrivelse
MetadataShow full item record
This report describes a model for cost-effective force structuring analysis using mathematical programming and optimization with game-theoretic utility measure. Force components including C2 and sensor/information are included in a multi-scenario setting to analyse a best overall expected force structure within given budget cost constraints. Each scenario uses a graph with nodes and connections having characteristics describing terraintypes, axis capacities and distances. Components manoeuvre and Lanchester type engagements take place in graph nodes using effectiveness data from lower level models. A main challenge for model development is the representation of strategies for application of game-theoretic methods with maximin solutions. Detailed user specified strategies, node restricted strategies as well as model optimized strategies are possible. A special algorithm for alternating side optimization is developed for approximate maximin mixed strategy solutions in an iterative loop for each scenario. Maximum expectation over all scenarios gives the reulting force structure. The model is developed using the AMPL mathematical programming tool with solver MINOS5.5 and a special heuristic solver HOTDOG also developed as part of the model.