Extensions to Linear Scheduling Optimization

Abstract
A two‐state‐variable, N‐stage dynamic programming formulation of the linear scheduling problem is presented. The state variables are vectors. For any one activity, the first state variable represents a set of possible durations required to complete work at each of the locations. Likewise, for any one activity, the second state variable represents a set of possible interrupt durations between work performed at adjacent locations. Choices of activity duration and interrupt duration vectors are considered for each of the activities involved in a project. The problem is formulated within a conventional dynamic programming framework with the objective of minimizing the overall project duration. The methodology accounts for several of the realities of repetitive construction, including generalized precedence relationships and the ability to treat a variety of work continuity constraints. In addition, a sensitivity analysis procedure is described which permits the identification of near‐optimal solutions, providing the user with schedule alternatives that might suit additional non‐quantifiable criteria better. The Selinger bridge construction example is used to illustrate application of the two‐state‐variable formulation and sensitivity analysis procedure.

This publication has 6 references indexed in Scilit: