A Heuristic Based on Matching and Bin Packing精选.pptVIP

Order-Fulfillment Rules Weight Constraints Internal properties the crane surface properties maximum allowble slab weights of every order the maximum weight of a slab that includes both and is The Slab Design Problem in the Steel Industry 10.3 A Heuristic Based on Matching and Bin Packing We developed a heuristic to produce solutions of good quality within a few minutes. The main building blocks of the heuristic are a nonbipartite matching step and a bin packing step. Given the order book, we construct an order-compatibility graph G = (V,E) as follows: : node edge e = : the maximum assigned weight of slab : the maximum weight of slab Step 1: Nonbipartite Matching The Slab Design Problem in the Steel Industry 7X1=7 1 7 O2 23 8X2=16 2 8 O1 Total weight Weight Number of PUs PUs 7X2=14 2 7 O2 23 9X1=9 1 9 O1 Total weight Weight Number of PUs PUs Among all such assignments, we choose the one with the maximum assigned weight as the weight of the edge joining the two orders. 7X1=7 1 7 O2 25 9X2=18 2 9 O1 Total weight Weight Number of PUs PUs O1: 8≤PUs ≤9 O2: 6≤PUs ≤7 The Slab Design Problem in the Steel Industry the maximum weight of a slab that includes both O^ and O^ is W = m.m{W^^^^W^^^). However, finding the maximum assigned weight is, in general, a nontrivial problem since it requires solving the following problem (P): The Slab Design Problem in the Steel Industry A maximum matching, M, in C identifies a set of order pairs, S, such that (1) each order pair is compatible (i.e., the two orders in the pair can be packed on the same slab) and (2) the total assigned weight for S is the maximum among all sets of order pairs with the property that a set contains each order at most once. A maximum weight matching The matching M provides us with |M| slabs each with either one or two orders. Each edge (i, j ) of the matching corresponds to one slab. This slab c