Ravi Sethi

The Complexity of Flowshop and Jobshop Scheduling

NP-complete problems form an extensive equivalence class of combinatorial problems for which no nonenumerative algorithms are known. Our first result shows that determining a shortest-length schedule in an m-machine flowshop is NP-complete for m ≥ 3. For m = 2, there is an efficient algorithm for finding such schedules. The second result shows that determining a minimum mean-flow-time schedule in an m-machine flowshop is NP-complete for every m ≥ 2. Finally we show that the shortest-length schedule problem for an m-machine jobshop is NP-complete for every m ≥ 2. Our results are strong in that they hold whether the problem size is measured by number of tasks, number of bits required to express the task lengths, or by the sum of the task lengths.

Scheduling independent tasks to reduce mean finishing time

Sequencing to minimize mean finishing time (or mean time in system) is not only desirable to the user, but it also tends to minimize at each point in time the storage required to hold incomplete tasks. In this paper a deterministic model of independent tasks is introduced and new results are derived which extend and generalize the algorithms known for minimizing mean finishing time. In addition to presenting and analyzing new algorithms it is shown that the most general mean-finishing-time problem for independent tasks is polynomial complete, and hence unlikely to admit of a non-enumerative solution.

Variations on the Common Subexpression Problem

Let G be a directed graph such that for each vertex v in G, the successors of v are ordered Let C be any equivalence relation on the vertices of G. The congruence closure C* of C is the finest equivalence relation containing C and such that any two vertices having corresponding successors equivalent under C* are themselves equivalent under C* Efficient algorithms are described for computing congruence closures in the general case and in the following two special cases. 0) G under C* is acyclic, and (it) G is acychc and C identifies a single pair of vertices. The use of these algorithms to test expression eqmvalence (a problem central to program verification) and to test losslessness of joins in relational databases is described

From regular expressions to deterministic automata

The main theorem allows an elegant algorithm to be refined into an efficient one. The elegant algorithm for constructing a finite automaton from a regular expression is based on ‘derivatives of’ regular expressions; the efficient algorithm is based on ‘marking of’ regular expressions. Derivatives of regular expressions correspond to state transitions in finite automata. When a finite automaton makes a transition under input symbol a, a leading a is stripped from the remaining input. Correspondingly, if the input string is generated by a regular expression E, then the derivative of E by a generates the remaining input after a leading a is stripped. Brzozowski (1964) used derivatives to construct finite automata; the state for expression E has a transition under a to the state for the derivative of E by a. This approach extends to regular expressions with new operators, including intersection and complement; however, explicit computation of derivatives can be expensive. Marking of regular expressions yields an expression with distinct input symbols. Following McNaughton and Yamada (1960), we attach subscripts to each input symbol in an expression; (ab + b)∗ba becomes (a1b2+b3)∗b4a5. Conceptually, the efficient algorithm constructs an automaton for the marked expression. The marks on the transitions are then erased, resulting in a nondeterministic automaton for the original unmarked expression. This approach works for the usual operations of union, concatenation, and iteration; however, intersection and complement cannot be handled because marking and unmarking do not preserve the languages generated by regular expressions with these operators.

The Generation of Optimal Code for Arithmetic Expressions

The problem of evaluating arithmetic expressions on a machine with N ≥ 1 general purpose registers is considered. It is initially assumed that no algebraic laws apply to the operators and operands in the expression. An algorithm for evaluation of expressions under this assumption is proposed, and it is shown to take the shortest possible number of instructions. It is then assumed that certain operators are commutative or both commutative and associative. In this case a procedure is given for finding an expression equivalent to a given one and having the shortest possible evaluation sequence. It is then shown that the algorithms presented here also minimize the number of storage references in the evaluation.

Complete Register Allocation Problems

The search for efficient register allocation algorithms dates back to the time of the first FORTRAN compiler for the IBM 704. The model we use in this paper is a single processor with an arbitraril...

A Level Algorithm for Preemptive Scheduling

Muntz and Coffman give a level algorithm that constructs optimal preemptive schedules on identical processors when the task system is a tree or when there are only two processors available. Their algorithm is adapted here to handle processors of different speeds. The new algorithm is optimal for independent tasks on any number of processors and for arbitrary task systems on two processors, but not on three or more processors, even for trees. By taking the algorithm as a heuristic on m processors and using the ratio of the lengths of the constructed and optimal schedules as a measure, an upper bound on its performance is derived in terms of the speeds of the processors. It is further shown that 1.23√m is an upper bound over all possible processor speeds and that the 1.23√m bound can be improved at most by a constant factor, by giving an example of a system for which the bound 0.35√m can be approached asymptotically.

Computing Sequences with Addition Chains

Given a sequence

Storage requirements for deterministic polynomialtime recognizable languages

n_1 , \cdots ,n_m

Code Generation for a One-Register Machine

of positive integers, what is the smallest number of additions needed to compute all m integers starting with 1? This generalization of the addition chain (