A Multi-Exchange Neighborhood for Minimum Makespan Parallel Machine Scheduling Problems

Process reassignment with reduced migration cost in grid load rebalancing

“…Each local search step (the "while" loop in lines 2-17) is refined into multiple sub-steps (the "for" loop in lines [5][6][7][8][9][10][11][12][13][14]. The different restrictions on acceptable migration cost of a process movement (thresholds on acceptable migration cost) are defined in the sub-steps.…”

“…Most local search-based heuristics in the literature use 1-move or 1-swap or their variations or combinations as the neighborhood structure [11][21][8] [9]. But they use local search solely without any meta-heuristic strategy (such as the guided local search in our algorithm).…”

Lower bounds and heuristic algorithms for the ki-partitioning problem

“…Hübscher and Glover [18] proposed a tabu search algorithm, FatemiGhomi and Jolai-Ghazvini [7] a local search approach based on 2-exchanges. Mendes et al [23] compared a tabu search approach with a memetic algorithm, while Frangioni, Necciari and Scutellà [8] presented a local search algorithm based on multiple exchanges within large neighborhoods.…”

Transforming Mathematical Models Using Declarative Reformulation Rules

“…A view based on complexity theory was proposed in [2], but since it requires a polynomial time mapping between the problems it already cuts off a number of well-known reformulation techniques where the mapping is pseudo-polynomial [6] or even exponential in theory [3,5,8], but quite effective in practice. Limited to MIP problems, general ideas based on variable redefinition were proposed by [13,14], without finding wide application due to its complexity, remaining unknown (or unused) by the "average" user.…”