Skills-Based Staffing Optimization via Iterative Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining staffing levels and scheduling in call centers, both skills-based and non-skills based, face challenges in efficiently meeting service performance targets due to limitations in existing simulation and optimization models, which fail to accurately account for multi-skill agents and customer abandonments.
Innovation Solution
A computer-implemented method and system that uses simulation and Integer Linear Programming (ILP) to determine staffing levels and schedule agents, iteratively unscheduling and rescheduling skill groups while simulating contact center environments to meet service level targets, incorporating customer abandonments and multi-skill agent capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional simulation and optimization models are used for staffing determination, then the modeling process is simpler, but the accuracy in meeting service performance targets deteriorates due to failure to account for multi-skill agents and customer abandonments
Solution Approach 1:
The patent segments the call center workforce into multiple skill groups, where each group possesses specific skills. This segmentation allows the simulation model to accurately track which agents can handle which types of calls, thereby improving the precision of service level predictions while accounting for multi-skill capabilities. The segmentation of agents by skill set enables more realistic modeling of call routing and agent availability.
Solution Approach 2:
The patent implements feedback mechanisms by iteratively simulating call center operations with proposed staffing levels and schedules, then using the simulation results to refine and adjust the staffing determination. The system continuously monitors service level performance metrics from the simulation and adjusts staffing recommendations accordingly, creating a closed-loop system that improves accuracy through iterative refinement while managing model complexity.
2Reliability
If iterative unscheduling and rescheduling of skill groups is performed, then service level targets are met more accurately, but the computational time and processing complexity increase
Solution Approach 1:
The patent performs preliminary actions by pre-calculating and storing skill group characteristics, call volume forecasts, and service level requirements before the iterative optimization process begins. This preliminary preparation reduces the computational burden during iterative unscheduling and rescheduling operations, as the system can quickly evaluate alternative staffing configurations without recalculating fundamental parameters, thereby reducing computational time while maintaining service level accuracy.
Solution Approach 2:
The patent applies partial action by focusing the iterative optimization process on specific skill groups that have the greatest impact on service level performance, rather than uniformly adjusting all skill groups in each iteration. This selective approach reduces the number of iterations required to converge on an optimal staffing solution, thereby reducing computational time while still achieving accurate service level target achievement.
3Measurement precision
If multi-skill agent capabilities are incorporated into the model, then staffing optimization accuracy improves, but the model complexity and difficulty of scheduling increase
Solution Approach 1:
The patent implements universality by creating a flexible simulation and optimization framework that can handle multiple skill types and agent capabilities within a unified model structure. The system uses universal data structures and algorithms that can accommodate any number of skill groups and their varying capabilities, allowing the model to scale from simple single-skill scenarios to complex multi-skill environments without requiring fundamentally different modeling approaches, thereby managing complexity while improving accuracy.
Solution Approach 2:
The patent introduces intermediary elements such as skill matrices and capability mappings that serve as mediators between the complex multi-skill agent capabilities and the staffing optimization algorithm. These intermediary structures organize and standardize the representation of multi-skill information, making it easier for the optimization model to process and utilize multi-skill agent data, thereby reducing the perceived complexity while maintaining high staffing determination accuracy.
Data Source
AI summary
An invention for scheduling employees in a skills-based contact routing environment in which each employee has one or more skills and belongs to a skill group. The invention includes an initialization step and an iterative step. In the initialization step, the invention generates initial staffing levels and initial agent schedules for each skill group. In the iterative step, a skill group is selected and unscheduled while keeping agents in other skill groups on their current schedules, agent requirements for the skill group selected are updated using a staffing model that applies contact routing rules and includes the number of agents scheduled for work in other skill groups, and agents are rescheduled in the selected skill group. In each iteration each skill group is unscheduled, agent requirements are updated, and agents are rescheduled once. Iterations continue until one or more stopping criteria are satisfied.