Multi-Objective Optimization — Hands-on

Learn practical multi-objective optimization (MOO) for medicinal chemistry: balance potency, selectivity, ADMET and developability under synthesis and building-block constraints. You will use Pareto fronts, desirability functions, evolutionary algorithms (NSGA-II/III) , and Bayesian multi-objective optimization to generate diverse, robust candidate sets with transparent trade-off rationale and reproducible reports.

Multi-Objective Optimization (Potency / Selectivity / ADMET)

Help Desk · WhatsApp

Session Index

Session 1 — Objectives, Constraints & Data Session 2 — Pareto & Desirability Approaches Session 3 — NSGA & Bayesian MOO Session 4 — Mini Capstone: Balanced Set

Session 1

Fee: Rs 21800

Objectives, Constraints & Data Setup

Define objectives & guardrails

potency & selectivity indices ADMET windows (tPSA/logD/etc.) project-specific hard/soft constraints

Data & models for MOO

QSAR surrogates & uncertainty applicability domain flags series-aware validation

Synthesis & availability constraints

BB coverage & retrosyn feasibility cost/lead-time heuristics IP/alert filters (lite)

Session 2

Fee: Rs 24800

Pareto Fronts & Desirability Functions

Pareto analysis & trade-offs

dominance, crowding distance knee points & frontier shape stability across seeds

Scalarization & desirability

weighted sums & Tchebycheff Harrington/Goal programming penalties & constraint handling

Diversity & coverage

scaffold/chemotype quotas medoid/cluster picks batching for synthesis/HTE

Session 3

Fee: Rs 27800

NSGA-II/III & Bayesian Multi-Objective BO

Evolutionary MOO (NSGA-II/III)

population, crossover, mutation constraint & penalty tuning runtime & convergence checks

Bayesian multi-objective optimization

GP/surrogates & UQ EHVI/ParEGO/TS policies batch BO & lab budgets

Synthesis-aware candidate selection

BB/retrosyn filters cost/lead-time weighting risk bands & alternates

Session 4

Fee: Rs 31800

Mini Capstone: Balanced Candidate Set

Build a potency–selectivity–ADMET MOO pipeline

Theory + Practical

Frontier analysis & diversity-batched picks

knee-point selection scaffold quotas synthesis feasibility overlay

Deliverables

Pareto plots & report (PDF/HTML) ranked candidate CSV configs/notebooks for reruns

PDF

Multi-Objective Optimization (Potency/Selectivity/ADMET) | Pareto, Desirability & Bayesian MOO

Multi-Objective Optimization — Hands-on