Abstract
This paper discusses the problem of assigning \(N\) streams of requests (clients) to \(M\) related server machines with the objective to minimize the sum of worst-case processing times. The completion time of a batch of requests is measured as a sum of weights of the subset of clients which share a single machine. Such problem can be seen as minimizing the sum of total weights of blocks of \(M\)-partition, each multiplied by the cardinality of a block. We prove that this problem can be solved in polynomial time for any fixed \(M\) and present an efficient backward induction algorithm.
Highlights
Consider the following problem of assigning requests to processing servers in distributed computing system
Parameter h j is the time of single request processing on jth server
The class of quadratic semi-assignment problem instances presented in this paper models the task of assigning streams of requests to a group of related server machines
Summary
Consider the following problem of assigning requests to processing servers in distributed computing system. Parameter h j is the time of single request processing on jth server. The processing requirement after assigning ith client to jth machine is ci j = wi h j. One specific class of polynomially solvable instances of QSAP has been characterized in [9], where its applications in flight scheduling are discussed. In this paper another class of QSAP instances of practical importance is identified, which can be solved efficiently, under certain fixed-parameter assumption. When M is considered as a part of the input, problem (1)–(3) is NP-hard [3]
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have