Balanced number partitioning

A common special case called two-way balanced partitioning is when there should be two subsets (m = 2). The two subsets should contain floor(n/2) and ceiling(n/2) items. It is a variant of the partition problem. It is NP-hard to decide whether there exists a partition in which the sums in the two subsets are equal; see [4] problem [SP12]. There are many algorithms that aim to find a balanced partition in which the sum is as nearly-equal as possible.

• Coffman, Frederickson and Lueker[5] present a restricted version of the LPT algorithm (called RLPT), in which inputs are assigned in pairs. When inputs are uniformly-distributed random variables, the expected largest sum of RLPT is exactly ${\displaystyle {\frac {n}{4}}+{\frac {1}{2n+2}}}$. The expected work-difference (difference between largest and smallest sum) is ${\displaystyle \Theta (1/n)}$.[2]
• Lueker[6] presents a variant of the LDM algorithm (called the pairwise differencing method (PDM)). Its expected work-difference is ${\displaystyle \Theta (1/n)}$.
• Tsai[2] presents an algorithm called Restricted Largest Difference (RLD). Its work-difference is ${\displaystyle O(\log n/n^{2})}$ almost surely.
• Yakir[7] presents a balanced variant of the LDM algorithm for m = 2, called BLDM. Its expected work-difference is ${\displaystyle n^{-\Theta (\log n)}}$.
• Mertens[8] presents a complete anytime algorithm for balanced two-way partitioning. It combines the BLDM algorithm with the complete-Karmarkar–Karp algorithm.

Another special case called 3-partitioning is when the number of items in each subset should be at most 3 (k = 3). Deciding whether there exists such a partition with equal sums is exactly the 3-partition problem, which is known to be strongly NP-hard. There are approximation algorithms that aim to find a partition in which the sum is as nearly-equal as possible.

• Kellerer and Woeginger[9] adapt the LPT algorithm to triplet partitioning (where there are at most 3*m items, and each subset should contain at most 3 items). Their algorithm is called modified-LPT or MLPT. It orders the items from large to small, and puts each item in turn into the bin with the smallest sum among those bins that contain fewer than 3 items. They show that the MLPT algorithm attains at most ${\displaystyle {\frac {4m-1}{3m}}}$ of the minimum largest sum, which is the same approximation ratio that LPT attains for the unconstrained problem. The bound is tight for MLPT.
• Chen, He and Lin[10] show that, for the same problem, MLPT attains at least ${\displaystyle {\frac {3m-1}{4m-2}}}$ of the maximum smallest sum, which is again the same ratio that LPT attains for the unconstrained problem.
• Kellerer and Kotov[11] present a different algorithm (for the case with exactly 3*m items), that attains at most ${\displaystyle 7/6}$ of the minimum largest sum.

A more general case, called k-partitioning,[12] is when the number of items in each subset should be at most k, where k can be any positive integer..

• Babel, Kellerer and Kotov[12] study a variant in which there are k×m items (for some integer k), and each of the m sets must contain exactly k items. They present several heuristic algorithms for approximating the minimum largest sum:
  • Folding algorithm: optimal for m = 2, and in general has tight approximation ratio ${\displaystyle 2-{\frac {1}{m}}}$.
  • Exchange algorithm: tight approximation ratio ${\displaystyle 2-{\frac {2}{m+1}}}$. It is not known if it runs in polynomial time.
  • Primal-dual algorithm (a combination of LPT and MultiFit): approximation ratio at most ${\displaystyle 4/3}$. It is tight for k = 4 when m is sufficiently large; the precise lower bound is ${\displaystyle {\frac {4m}{3m+1}}}$).[12]^: Thm.11
  • The also conjecture that Modified-LPT has an approximation ratio ${\displaystyle {\frac {4m-1}{3m}}}$ . Currently, this conjecture is known to be true only for k = 3.[9] For k > 3, it is known that its approximation ratio is at most 2.[3]
• Michiels, Korst, Aarts, van Leeuwen[13] and Spieksma[14] study a variant in which each of the m sets must contain either ceiling(n/m) or floor(n/m) items (so k = ceiling(n/m)). They extend the Balanced-LDM (BLDM) from m=2 to general m. The generalized algorithm runs in time ${\displaystyle O(n\log n)}$. They prove that its approximation ratio for the minimum largest sum is exactly 4/3 for k = 3, 19/12 for k = 4, 103/60 for k = 5, 643/360 for k = 6, and 4603/2520 for k = 7. The ratios were found by solving a mixed integer linear program. In general (for any k), the approximation ratio is at least ${\displaystyle 2-\sum _{j=0}^{k-1}{\frac {j!}{k!}}}$ and at most ${\displaystyle 2-{\frac {1}{k-1}}}$. The exact MILP results for 3,4,5,6,7 correspond to the lower bound. For k>7, no exact results are known, but the difference between the lower and upper bound is less than 0.3%. When the parameter is the number of subsets (m), the approximation ratio is exactly ${\displaystyle 2-{\frac {1}{m}}}$.
• Zhang, Mouratidis and Pang[1] show that BLDM might yield partitions with a high work-difference (difference between highest and lowest sum), both when the inputs are distributed uniformly, and when their distribution is skewed. They propose two alternative heuristics: LRM reduces the work-difference to 1/3 the work-difference of BLDM when the distribution is uniform; Meld reduces the work-difference when the distribution is skewed. A hybrid algorithm combines BLDM, LRM and Meld and adapts dynamically to different data distributions.
• When k is fixed, a PTAS of Hochbaum and Shmoys[15] can be used for balanced partitioning.[14] When k is part of the input, no PTAS is currently known.[14]
• Dell'Amico and Martello[3] study the problem of minimizing the largest sum when the number of items in all sets is at most k. They show that the linear-program relaxation of this variant has the same optimal value as the LP relaxation of the unconstrained variant. The expression ${\displaystyle \max((\sum x_{i})/m,x_{1})}$, where x_i are the inputs ordered from large to small, is a lower bound for the optimal largest-sum, and its worst-case ratio is 1/2 in both variants. The improved expression ${\displaystyle \max((\sum x_{i})/m,x_{1},x_{k}+x_{m+1})}$ has worst-case ratio 2/3 in the unconstrained variant and 1/2 in the constrained variant. The approximation ratio of the modified list scheduling is 1/2 for the unconstrained variant, but it is 0 for the constrained variant (it can be arbitrarily bad). The approximation ratio of the modified LPT algorithm is at most 2. They also show that the lower bound of [12] has a tight worst-case performance ratio of 3/4, and that their PD algorithm has a tight performance ratio of 4/3 (when m is sufficiently large).
• He, Tan, Zhu and Yao[16] consider the problem of maximizing the smallest sum. They show that the FOLDING algorithm has tight approximation ratio ${\displaystyle \max \left({\frac {2}{k}},{\frac {1}{m}}\right)}$. They present a new algorithm, HARMONIC1, with worst-case ratio at least ${\displaystyle \max \left({\frac {1}{k}},{\frac {1}{\lceil \sum _{i=1}^{m}{\frac {1}{i}}\rceil +1}}\right)}$. Both these algorithms are ordinal – they partition the items based only on the order between them rather than their exact values. They prove that any ordinal algorithm has ratio at most ${\displaystyle O(1/\ln {m})}$for maximizing the smallest sum. This indicates that HARMONIC1 is asymptotically optimal. For any fixed k, any ordinal algorithm has ratio at most the smallest root of the equation ${\displaystyle \sum _{i=1}^{m}\left\lfloor \left\lfloor {\frac {k+i-1}{i}}\right\rfloor x\right\rfloor =k}$. When k tends to infinity, this upper bound approaches 0.

The cardinality constraints can be generalized by allowing a different constraint on each subset. This variant is introduced in the "open problems" section of,[12] who call the k_i-partitioning problem. He, Tan, Zhu and Yao[16] present an algorithm called HARMONIC2 for maximizing the smallest sum with different cardinality constraints. They prove that its worst-case ratio is at least ${\displaystyle \max \left({\frac {1}{k_{m}}},{\frac {k_{1}}{k_{m}}}{\frac {1}{\left\lceil \sum _{i=1}^{m}{\frac {1}{i}}\right\rceil +1}}\right)}$.

Another generalization of cardinality constraints is as follows. The input items are partitioned into k categories. For each category h, there is a capacity constraint k_h. Each of the m subsets may contain at most k_h items from category h. In other words: all m subsets should be independent set of a particular partition matroid. Two special cases of this problem have been studied.

Matroid-constrained number partitioning is a generalization in which a fixed matroid is given as a parameter, and each of the m subsets should be an independent set or a base of this matroid.

• Cardinality constraints are special cases of matroid constraints in which the matroid is a uniform matroid.
• Categorized cardinality constraints are a special case in which the matroid is a partition matroid.