The efficient algorithm for optimal capacity expansion of satellite links in mobile networks

A deterministic expansion model for sizing of N satellite links in mobile networks during exploitation is being developed. The model is applied to short- or medium-term planning with finite number of discrete time periods. In the planning process it can help the Land Earth Station Operator (LESO) to find the optimal capacity-expansion policy for new constructions and conversions of channel equipment that minimizes the total cost. The model for the capacity expansion problem with shortages (CEPS) allows capacity conversions from one satellite link to another in both directions, but limitations can be imposed. Furthermore, upper bounds on shortages and idle capacities can be imposed. The traffic demand increments for additional capacity in any time period and for any link can be negative. All cost functions are assumed to be piecewise concave and non-decreasing away from zero. Using a network flow approach the paper develops an efficient dynamic-programming algorithm to minimize the total expansion cost. An heuristic algorithm is tested on many examples, and its efficiency is compared with an algorithm based on an exact approach.