TR155: Formalizing Construction Sequence Constraints for the Rapid Generation of Scheduling Alternatives

Construction planners today use CPM-based schedules to represent the planned logical sequence of activities to perform a project. A construction schedule typically represents the sequence of multiple trades that perform individual work while sharing common workspaces or resources. During the course of a project, planners frequently modify activity sequences to meet changing project demands. More than one sequencing alternative can exist, and planners need to develop and evaluate alternative sequences to make well-informed decisions.
When developing sequencing alternatives, planners need to understand the “role” an activity plays for following activities. They also need to assess the “status” of activities, i.e., whether activities may or may not be delayed. Planners infer the role and status of activities by conceptually classifying the rationale for constraints between activities with respect to their role and flexibility.

However, the current CPM framework only distinguishes the temporal aspects of constraints and only distinguishes the time-criticality of activities. Consequently, determining the role and status of activities can only be performed in the planner’s minds. Correspondingly, developing sequencing alternatives using today’s CPM-based scheduling tools is an error-prone and time-consuming process.

Thus, the goals of this research were threefold: 1) develop a representation of sequencing rationale that enable planners to describe their rationale for constraints and the classifications they make for different types of sequencing rationale, 2) develop a mechanism that leverages the representation to infer the role and status of activities automatically, and 3) develop a process that supports planners in utilizing the representation and mechanism to develop sequencing alternatives correctly and rapidly.

Accomplishing these goals posed unique challenges. The representation needs to model a classification schema that correctly classifies the different types of specific constraints (e.g., damaged by, protected by, etc.) in construction schedules with respect to their role and flexibility. The mechanism needs to identify unique paths between activities in a CPM network and correctly classify activities based on the role and flexibility of the individual constraints in these paths. The process needs to model generically how planners use the role and status of activities to identify which activities to delay and to prioritize activities when developing sequencing alternatives.

I addressed these challenges by formalizing (1) an ontology that models the conceptual classification planners make for construction sequencing rationale, (2) a “classification” mechanism that uses a network chain search algorithm and inference rules to infer the role and status of activities automatically given a CPM network schedule for which sequencing rationale has been explicitly represented using the ontology, and (3) a formal process that integrates the classification mechanism and ontology to guide planners in developing sequencing alternatives correctly and rapidly.

I demonstrated the power and generality of the formalizations by performing three retrospective test cases and one charrette test using prototype software, CLCPM. The retrospective cases validated that the ontology correctly classifies the different types of constraints that exist in construction schedules. The cases also validated that the classification mechanism and formal process could be used to identify and re-sequence activities correctly for different types of construction work. The charrette test demonstrates that planners identify and re- sequence activities more accurately and consistently using CLCPM than with conventional scheduling tools.

Practically, the formalizations provide an environment that provides planners with a better basis to make re-sequencing decisions. The environment enables multiple project members to communicate and discuss the logic of activity sequences, better understand the individual role and impact between project members, and evaluate multiple sequencing scenarios more quickly and consistently within group settings.